Vue normale

Two Kalibr-carrying frigates, a missile ship, a patrol vessel: Ukraine hit four Russian warships in one night at Novorossiysk

12 août 2026 à 16:38

'Admiral Makarov' in 2020

Ukraine reached the port where Russia's Black Sea Fleet took shelter and hit four of its warships. On the night of 12 August 2026, in a joint operation, Ukraine's Defense Forces struck the naval base at Novorossiysk in Russia's Krasnodar Krai, the General Staff reported.

Novorossiysk became Russia's main Black Sea naval base after Ukrainian strikes and sea drones drove the fleet out of occupied Sevastopol in Crimea, so reaching its warships there strikes at the refuge that was supposed to be safe.

The ships that were hit

The four vessels are among the fleet's more capable surface combatants. The two Project 11356 frigates, Admiral Makarov and Admiral Essen, are among the newest and most important surface ships Russia has kept in the Black Sea, both capable of launching Kalibr cruise missiles at Ukraine.

The Buyan-M is another Kalibr shooter. The Project 21631 small missile ship carries the same cruise missiles Russia fires at Ukrainian cities, and the Vasily Bykov is a patrol ship of the class Ukraine has struck before, so the strike damaged both major missile-carrying combatants and a patrol vessel in one night.

The overall damage at the base is still being assessed.

The fleet has no safe harbor left

The strike extends a campaign that has steadily pushed Russia's fleet eastward and out of action. Ukraine drove the Black Sea Fleet from Sevastopol without a navy of its own, using drones and missiles, forcing Russia to pull its ships back to Novorossiysk, farther from Ukraine but not out of reach.

This means that Russia holds no waters it can use and now no harbor where its ships are safe. Ukraine's Defense Forces said they consistently reduce Russia's capacity to continue its armed aggression, and damaging four warships at their fallback base is the naval form of that pressure.

Russia starves most Ukrainian POWs, then “fattens” a few for exchange so they look normal, Ukrainian official says

12 août 2026 à 16:28

Ukranian POW returning from Russian captivity, 5 February 2026. Photo: Coordination Headquarters for the Treatment of Prisoners of War

Russia is forcing Ukrainian prisoners into its own army and starving others while "fattening" a few for show, Andrii Pasternak of the SBU's Joint Center for coordinating the search and release of prisoners says. Ukrainian intelligence also has information about Moscow creating battalions from Ukrainian prisoners of war (POWs), UkrInform reports.

As of May 2026, approximately 7,000 Ukrainian service members were held in Russian captivity, while the number of civilian detainees is estimated at between 10,000 and 20,000, according to Vice Speaker of the Ukrainian Parliament Olena Kondratiuk.

Russia is forming units from Ukrainian prisoners

Pasternak described the POW battalions as confirmed and monitored.

"We know about such battalions. Together with our colleagues from foreign intelligence and military intelligence, we track and record this. Such facts exist," he stated.

Forcing prisoners of war to serve in the captor's forces is itself a war crime. The Geneva Conventions prohibit compelling a POW to fight against their own country, and turning captured Ukrainian soldiers into Russian units, if confirmed at scale, would be a grave violation on top of the coercion required to make men fight the country they were defending.

Russia feeds up some prisoners and starves others

The coordinator also pointed to a practice of staging appearances before exchanges. Knowing whom it will hand over, Russia brings those prisoners, including civilians, into "normal form" through a process he called "fattening," drawing on testimony from released Ukrainian defenders and returned civilians.

The starvation behind it is deliberate and selective. Pasternak characterized it as an element of torture: some prisoners are deliberately not fed, while others are fed up so they look more or less normal for the swap. Ukraine's negotiators note at every meeting that the two sides' prisoners differ starkly, he added. The returning Ukrainians are almost all very thin, with only a few in decent condition, while the Russian POWs Ukraine hands back are genuinely well-fed.

  • ✇Euromaidan Press
  • Russia builds 3,000 jet drones monthly to counter Ukrainian interceptors, HUR says
    Russia has ceased producing its Geran-3 jet-powered attack drone, Ukraine’s Military Intelligence (HUR) told the Ukrainian defense outlet Militarnyi. The agency estimates Russian industry now makes about 3,000 newer Geran-4 and Geran-5 drones per month. Russia is shifting toward faster jet drones as Ukraine’s low-cost interceptors have become more effective against slower Shaheds. The Geran-4 and Geran-5 leave defenders less time to react, requiring faster interceptors
     

Russia builds 3,000 jet drones monthly to counter Ukrainian interceptors, HUR says

12 août 2026 à 14:02

3D model of a Russian Geran-4 jet-powered attack drone, shown from above with its tracker and navigation unit listed beside it.

Russia has ceased producing its Geran-3 jet-powered attack drone, Ukraine’s Military Intelligence (HUR) told the Ukrainian defense outlet Militarnyi. The agency estimates Russian industry now makes about 3,000 newer Geran-4 and Geran-5 drones per month.

Russia is shifting toward faster jet drones as Ukraine’s low-cost interceptors have become more effective against slower Shaheds. The Geran-4 and Geran-5 leave defenders less time to react, requiring faster interceptors or short-range air defenses. At the reported production rate, these newer Geran models, previously used in limited numbers, could become a regular feature of Russian attacks.

HUR said Russia has established serial assembly of the two newer models at the Alabuga complex in Tatarstan, with other Russian companies supplying components. The agency also said Russia’s Defense Ministry wants to expand Alabuga’s capacity. Construction began in May. Satellite imagery showed large-scale work by early July 2026, although the purpose of the new buildings remains unknown.

From Geran-3 to newer jet drones

Russia first used Geran-3 drones in large numbers during a combined nighttime attack on Ukraine on 30 July 2025. At least eight flew alongside propeller-driven Shaheds and decoys. The model retained the Geran-2 airframe but replaced its piston engine with a Chinese Telefly JT80 turbojet. HUR gave it a speed of 300–370 kilometers per hour and a range of up to 1,000 kilometers.

In May, then-Defense Ministry adviser Serhii “Flash” Beskrestnov said Ukrainian defenses stopped 95% of incoming Shaheds during the most recent documented strike and predicted Moscow would produce more jet-powered Geran variants.

The newer Geran-4 has a more robust airframe than the Geran-3 and is designed for high-speed flight and maneuvers. It can reach 500 kilometers per hour.

The larger Geran-5 carries a roughly 90-kilogram warhead and has a stated range of up to 1,000 kilometers. HUR has said Russia may equip it with R-73 air-to-air missiles, potentially threatening aircraft sent to intercept it.

The Chinese engines that Russia itself can't build

The Geran line combines Iranian designs, Russian assembly, and imported technology. HUR described the Geran-3 as a localized version of Iran’s Shahed-238. It found 45 foreign components inside, including parts from the United States, China, Switzerland, Germany, Britain, and Japan.

Russia depends on China for the turbojet engines in newer Geran models. HUR deputy chief Vadym Skibitskyi told RBC-Ukraine that Russia remains fully dependent on Chinese turbojet engines for the Geran-4 and Geran-5. Each costs about $35,000–$40,000, he said.

A Chinese XK-F358 mesh radio recovered from a Geran-5 also helps Russian drones resist jamming. Its manufacturer remains outside the US, EU, UK, and Ukrainian sanctions, according to a review by Euromaidan Press.

  • ✇Euromaidan Press
  • A Ukrainian railway engineer spied on Lyman’s defenses for Russia using Google Maps
    A Ukrainian court sentenced the former chief engineer of Lyman’s locomotive depot to 15 years in prison with property confiscation for helping Russia target the frontline city, Ukraine’s Security Service (SBU) said on 11 August. Investigators report he sent a Russian handler the locations of Ukrainian positions intended for guided-bomb and artillery strikes. Lyman blocks Russia’s northern route toward Sloviansk and Kramatorsk, which are part of Ukraine's fortress belt i
     

A Ukrainian railway engineer spied on Lyman’s defenses for Russia using Google Maps

12 août 2026 à 14:01

SBU officers in camouflage and tactical gear, seen from behind during an operation.

A Ukrainian court sentenced the former chief engineer of Lyman’s locomotive depot to 15 years in prison with property confiscation for helping Russia target the frontline city, Ukraine’s Security Service (SBU) said on 11 August. Investigators report he sent a Russian handler the locations of Ukrainian positions intended for guided-bomb and artillery strikes.

Lyman blocks Russia’s northern route toward Sloviansk and Kramatorsk, which are part of Ukraine's fortress belt in the Donbas. As attacks on supply lines intensify, information about Ukrainian positions can help Russia direct bombs and artillery. The case fits a broader pattern of Russian intelligence recruiting people inside Ukraine to locate targets. Russian services have used money and blackmail, including threats against recruits’ families, to secure their cooperation.

According to the SBU, the former engineer traveled around Lyman and its outskirts to record Ukrainian military bases. He marked possible targets on Google Maps, then sent screenshots to a handler in Russia through a messaging app.

The collaborator operated from basement

The agency said an acquaintance whom it described as a “DNR deputy” in occupied Donetsk Oblast connected him with Russian intelligence. The SBU said the man moved into an apartment building’s basement to shelter himself while directing Russian strikes on the community.

SBU officers detained him in Sloviansk in May 2026. The agency said a seized smartphone contained evidence of his reconnaissance and contacts with the Russian side.

The SBU did not identify the man, the court, or any specific strike enabled by his information.

How Russia turns locals into spotters 

The Lyman case involved an acquaintance of the former engineer, but Russian services also use money, blackmail, and online platforms to find recruits. The SBU says minors may initially receive simple assignments, such as taking photographs or painting graffiti, before being directed toward more serious sabotage.

Other recruits have allegedly answered Telegram offers of “easy money.” These include a Zaporizhzhia man later convicted of building bombs and a Dnipro suspect accused of activating more than 1,000 Starlink terminals for Russian forces.

Ukraine has prosecuted similar cases before. In November 2024, a court sentenced another Russian agent to 15 years for guiding missile and drone strikes in Zhytomyr Oblast.

More recent cases remain at the allegation stage. The SBU has accused a Kharkiv Oblast man of installing hidden cameras near military sites and a Poltava student of mapping energy facilities while filming air-defense units.

Russians work three shifts to build jet-powered missiles powered by a $16,000 Chinese engine from AliExpress

12 août 2026 à 12:50

Rocket S8000 Banderole

Russia is turning out its cheap Banderol drone-cruise missile at scale. Russia is capable of producing up to 80 jet-powered weapons a month, Ukraine's Defense Intelligence (HUR) told the defense outlet Militarnyi in response to a query.

The production figure sharpens a threat Ukraine has tracked as the Banderol expands. The missile has a relatively short range of about 500 kilometers, so Russia effectively launches it from Crimea at Odesa. It means Russia can sustain and grow the Banderol strikes on Ukraine's south, a weapon whose ground launcher Ukraine recently found could let Russia fire it without warning and in salvos.

The missile is fast, cheap, and short-range

The Banderol was developed and is made by the Russian company Kronshtadt, reaching a maximum speed of 620 to 650 kilometers per hour and cruising at 520 to 560, with a fragmentation-high-explosive warhead of 114.3 kilograms, of which 49.5 kilograms is explosive.

Its range dictates where it is used. At about 500 kilometers, the Banderol cannot reach far into Ukraine, which is why Russia fires it from occupied Crimea at Odesa, using the peninsula's proximity to strike the southern port city that has borne so much of Russia's missile campaign.

It has launched them at the Kyiv agglomeration to hit Rozetka, Ukraine's largest online retailer.

Banderol's jet propulsion allows speeds that Ukrainian mobile fire teams cannot match with standard ground-based machine guns. Countering the missile requires systems like the Gepard that can track fast-moving targets. 

The production runs on Western parts

The 80-a-month figure comes at a cost in effort and imports. To sustain the stated production rate, Russian enterprises work in three shifts, and to make the missiles, Russia uses Western components.

Ukraine's War & Sanctions portal documented over 20 foreign-made components and approximately 20 microchips of foreign origin from 30 different companies in the missile's construction, with a Chinese Swiwin SW800Pro jet engine, available on AliExpress for approximately $16,000, powering the weapon.

  • ✇Euromaidan Press
  • The Belarus border tunnels reached Latvia—guards turned back the 15 migrants who crawled underground
    Latvia has discovered its first known border tunnel used to enter the country illegally from Belarus, the country’s State Border Guard said on 11 August. Officers intercepted 15 people who had crossed through the underground passage and returned them to Belarus. Latvia, Lithuania, and Poland accuse Belarus and Russia of directing migrants toward their borders to retaliate for EU sanctions and strain countries supporting Ukraine. Minsk and Moscow deny orchestrating the c
     

The Belarus border tunnels reached Latvia—guards turned back the 15 migrants who crawled underground

12 août 2026 à 12:48

A flashlight illuminates the narrow entrance to a tunnel dug beneath the Latvia–Belarus border.

Latvia has discovered its first known border tunnel used to enter the country illegally from Belarus, the country’s State Border Guard said on 11 August. Officers intercepted 15 people who had crossed through the underground passage and returned them to Belarus.

Latvia, Lithuania, and Poland accuse Belarus and Russia of directing migrants toward their borders to retaliate for EU sanctions and strain countries supporting Ukraine. Minsk and Moscow deny orchestrating the crossings. Tunnels found in Lithuania and Latvia show organizers adapting to fortified borders, requiring Baltic governments to deploy more personnel and surveillance technology.

The tunnel was about 10 meters from Latvia's fence

Border guards found the group on 10 August in Skrudaliena parish, in Latvia’s southeastern Augšdaugava municipality, and briefed the public on the discovery the following day. The operation involved the State Police, a Lithuanian border-service dog handler, and a rapid-response unit, the Latvian Border Guard said.

While reconstructing the group’s route, officers found the passage near the Silene border post, about 10 meters from Latvia’s fence. The agency said it had received intelligence that a tunnel might be under construction, but had previously found no sign of a breach there.

Riga answered on 11 August 2026 by launching Operation Vilkatis, or “Werewolf,” scheduled to run through the end of August. The government says the operation will improve intelligence-sharing among border guards, police, the military, and security services. It will also expand aerial surveillance and target networks that transport and shelter migrants after they cross.

belarusian tunnel eu almost done lithuania drilled hole exactly wrong place—for diggers · post unfinished dug belarus toward lithuanian territory varėna district туннель ведущий из беларуси на территорию литвы vsat
Explore further

The Belarusian tunnel to the EU was almost done. Then Lithuania drilled a hole in exactly the wrong place—for the diggers

Latvia faced 87% of all crossing attempts along the EU-Belarus border by mid-July 2026

Belarus has weaponized migration since 2021, when EU governments accused Alyaksandr Lukashenka’s regime of bringing people from the Middle East and Africa to the bloc’s frontier in retaliation for sanctions. Belarus denied orchestrating the crisis.

Latvia reported more direct evidence in January 2026. Documents and phones taken from detained migrants linked Belarusian military and border personnel to organized crossings, according to Latvia’s armed forces and Border Guard.

By 13 July, Latvia had recorded 7,791 attempts over 111 consecutive days, 87% of all attempts along the EU-Belarus border by then. Lithuania later found an unfinished tunnel dug from Belarus. Latvia’s tunnel, unlike the unfinished Lithuanian passage, had already been used.

“They laid down their weapons, raised their hands”: Russia has executed at least 345 surrendered Ukrainian soldiers on the ground

12 août 2026 à 12:34

Orban's Gambit: why Russia and Hungary secretly moved Ukrainian POWs to Budapest

Russian forces have executed at least 345 Ukrainian soldiers who surrendered since the start of the full-scale war, Ukrainian Ombudsperson Dmytro Lubinets said at a press conference at Ukrinform.

The figure is the latest point on a documented and rising curve of a systematic war crime. Ukraine's prosecutors and international monitors have tracked these killings from 268 confirmed by July 2025 to over 330 by March 2026. Financial Times investigation found the executions appear to be systematic military policy directed from Russian command, with intercepted orders to "shoot prisoners on the spot."

The victims had surrendered

Lubinets described people killed after they had given up.

"It means that they laid down their weapons, came out with raised hands, lay on the ground, and showed they were no longer fighting," he said, and Russian soldiers examined them and confirmed they were unarmed.

Then the soldiers killed them.

"We have recorded cases where they fully demand that the clothes be removed. And after that, they force them to lie on the ground the same way, and shoot the Ukrainian soldiers at point-blank range," Lubinets added.

Killing a surrendered soldier is a war crime

Lubinets stressed the legal character of the act. The killing of soldiers who have surrendered is a distinct war crime, he said.

The prohibition is absolute. Under the Geneva Conventions, a soldier who lays down arms and surrenders is protected, and killing a prisoner of war is a grave breach of international law. Common Article 3 of the Third Geneva Convention prohibits "violence to life and person, in particular murder of all kinds" against those who have laid down their arms, with no battlefield circumstance that permits it, per the ICRC.

The count keeps rising and accountability lags

Ukraine documents each case for eventual prosecution, but most perpetrators are in Russia, beyond reach, and Russian courts and media have glorified the killings rather than punished them.

FPV drone repurposed to hunt Russian recon drones worked so well that Ukraine turned it into new purpose-built interceptor

12 août 2026 à 12:16

A TAF Kolibri-i10 interceptor. Source: TAF Industries

Ukraine has approved a new domestic interceptor and put it on the shelf for units to buy. Ukraine's Defense Ministry has codified the high-speed TAF Kolibri-i10 interceptor from TAF Industries, the company announced.

The approval adds one more weapon to Ukraine's fast-scaling interceptor arsenal and its state procurement pipeline. Ukraine now fields drone-hunters from dozens of makers to counter Russia's massed drone attacks. Units can order it directly through the state marketplace, the same DOT-Chain and Brave1 Market system that has turned battlefield-proven drones into off-the-shelf purchases.

The interceptor grew out of combat use

The Kolibri-i10 came from an improvised solution that worked. The idea grew from using the FPV drone Kolibri 10 to intercept Russian fixed-wing drones, and combat experience showed the approach was so effective that TAF built a new, specialized interceptor rather than modifying the existing model.

Its job is hunting Russian reconnaissance and strike drones. The Kolibri-i10 is designed to destroy Molniya, Orlan, and Zala drones, preventing attacks on Ukrainian positions and covering operations and rear-area sites.

"Field tests have already confirmed the effectiveness of this interceptor, so we have already begun scaling it," CEO Volodymyr Zinovskyi said.

The specs and the jam-resistance

The interceptor was rebuilt for speed and durability. It has a stronger frame, more powerful motors, faster propellers, and improved aerodynamics, reaching over 200 kilometers per hour, a range of up to 25 kilometers, more than 20 minutes aloft, and an altitude of up to 3,000 meters.

It is built to work under jamming. Manual control gives the operator full command at critical moments; digital day-and-night cameras enable operation in varied conditions. MilELRS firmware ensures stable operation in the presence of active electronic warfare. That jam-resistance is a recurring requirement for Ukrainian interceptors, since signal-dependent ones fail against the electronic warfare saturating the front.

The drone is available to order on Brave1 Market, priced at about $760 with a day camera and $1,550 with a night camera.

  • ✇Euromaidan Press
  • Russia steers its long-range drones from ground towers—Kyiv says it hit one in occupied Crimea
    Ukraine struck a Russian drone-control relay station and two other military targets in occupied Crimea and the occupied south overnight on 12 August, its General Staff said. Explosions rolled across the peninsula through the night, Sevastopol declared repeated air-raid alerts, and towns across occupied Crimea lost power, according to Ukrainian monitoring channels and Russian occupation authorities.  Russia launches its attack drones at Ukraine from sites inside Russia a
     

Russia steers its long-range drones from ground towers—Kyiv says it hit one in occupied Crimea

12 août 2026 à 12:14

russia steers its long-range drones ground towers—kyiv says hit one occupied crimea · post nasa firms satellite data shows fire about 6 km east olenivka chornomorske raion 12 2026 ukraine

Ukraine struck a Russian drone-control relay station and two other military targets in occupied Crimea and the occupied south overnight on 12 August, its General Staff said. Explosions rolled across the peninsula through the night, Sevastopol declared repeated air-raid alerts, and towns across occupied Crimea lost power, according to Ukrainian monitoring channels and Russian occupation authorities. 

Russia launches its attack drones at Ukraine from sites inside Russia and across occupied territory every day, and relay stations let it steer those drones over long distances. Occupied Crimea is the logistical spine of Russia's hold on southern Ukraine, the funnel through which fuel, ammunition, and reinforcements reach the front, and its fixed airfields, power plants, and control stations sit within reach of Ukraine's growing long-range arsenal.

What the General Staff says was hit

Ukraine's Defense Forces struck three targets overnight, the General Staff reported. The first was a ground relay station used to control Geran and Gerbera attack drones near Olenivka in occupied Crimea. The others were a Russian repair unit near Mariupol in Donetsk Oblast and a staging point for an enemy drone unit near Marivka in Luhansk Oblast.

Hitting these sites weakens Russia's ability to command its attack drones, keep its military vehicles running, and sustain its units, the General Staff said. 

Occupied Crimea has three villages named Olenivka, and the General Staff did not say which one it meant. The likeliest is in the Yevpatoria district, at the tip of the Tarkhankut peninsula—Crimea's westernmost cape and its closest point across the Black Sea to mainland Ukraine's Odesa Oblast, a frequent target of Russian drones launched from the peninsula.

Euromaidan Press could not independently verify the strikes and damage.

A photo shared by Robert “Madyar” Brovdi shows the Russian ferry “Maria” before it was struck in Ukrainian drone operations, with the damaged “Panagia” ferry visible to the right. Photo: Robert "Madyar" Brovdi
Explore further

Ukraine’s Crimea campaign could strike seven times more targets with faster funding, Brovdi says

Russia flies these drones through a chain of ground repeaters, often mounted on mobile-operator towers or similar tall structures, boost the control signal so operators inside Russia can steer Geran and Gerbera drones over Ukraine from hundreds of kilometers away. Kyiv has dismantled similar relay networks before, including one Russia had built on towers and rooftops in Belarus.

Explosions and blackouts across occupied Crimea

Sevastopol declared three air-raid alerts, and the Russian-installed governor Mikhail Razvozhaev claimed Russian forces downed more than 30 drones, with the heaviest activity over the Balaklava district and no one hurt.

The Ukrainian monitoring channel Krymsky Veter reported explosions in Feodosia and Kerch, and drones close to the Kacha and Belbek airfields. Residents reported air-defense activity near major cities.

The channel said explosions sounded near the Balaklava thermal power plant, a mobile gas-turbine station, and a substation in Sevastopol. Power cut out across parts of occupied Crimea, leaving Sevastopol, Simferopol, and several more cities and districts without electricity.

sea drones jet neptune missiles hit one russian port single night · post fire burns novorossiysk early hours 12 2026 пожежа у новоросійську ніч проти серпня фото exilenova+ ukraine news
Explore further

Sea drones, jet drones, and Neptune missiles hit one Russian port in a single night

After the attack, Krymsky Veter's monitoring group reported two large fires near Sevastopol, citing satellite data. One sat by a road junction near the village of Pervomaiske under Balaklava, where a bridge crosses the railway. The other burned near one of the local beaches. The group also logged an already-extinguished fire in the Inzhyr tract that had been reported the day before.

The same night, Ukraine also struck Russia's naval base at Novorossiysk, the fleet's main Black Sea base since late 2023.

Part of the campaign to isolate Crimea

The overnight strikes fit Ukraine's months-long effort to cut occupied Crimea off from the forces and supplies that sustain it. Ukrainian drones have severed road and rail links to the peninsula, left it rationing fuel under rolling blackouts, and pushed the campaign onto the sea by hunting the tankers that supply it. 

Ukraine finds American Nvidia Jetson micro-computer in Russia’s new missile that can’t reportedly tell truck from a bus with children

12 août 2026 à 12:03

The image shows a Russian missile. Source: HUR

An American AI computer has turned up inside Russia's newest cruise missile. Ukrainian specialists has found an Nvidia Jetson micro-computer in Russia's new S-71 "Monokhrom" cruise missile, the Defense Ministry's Main Intelligence Directorate (HUR) reports.

S-71M “Monokhrom is the first fully autonomous long-range strike system designed to engage not only stationary but also moving targets. The system is designed to independently decide whether to destroy a target based on threat priorities. A significant problem for such autonomous systems is determining exactly what they are attacking — a military truck or a bus carrying children.

The discovery is the latest evidence that Russia's newest weapons still run on Western technology. Ukraine has already documented hundreds of foreign components in Russian missiles and drones reaching Moscow despite sanctions.

The AI chip is the standout find

The use of the Nvidia Jetson may indicate the application of AI technologies in the missile. This capability would let the weapon process visual or sensor data onboard to identify targets or navigate, functions that depend on exactly the kind of compact AI computing it provides.

The chip sits among many foreign parts. In total, 35 electronic components used by Russia in the missile were identified, making Monokhrom another Russian weapon built substantially from foreign electronics.

The same parts show up across Russia's arsenal

HUR's analysis linked the missile to Russia's wider weapons program. The identified components included parts of a passive radar seeker that Russia has begun installing on Geran-2 drones to detect and strike Ukrainian air-defense systems and radars.

Other finds tied the missile to Russia's newest strike systems. They included a Chinese Honpho TS130C-01 camera from the jet-powered Geran-4 drone and components of the active homing seeker of the Kh-47M2 Kinzhal aeroballistic missile.

The finding sharpens the sanctions case

HUR framed the discovery as an argument for tighter enforcement.

"The discovery of an Nvidia Jetson in a new Russian missile once again demonstrates the need to strengthen sanctions pressure and coordinate the efforts of the civilized world," HUR said.

HUR specialists and Ukrainian research institutions continue to study Russian weapons and trace the origin of their electronic components to strengthen measures to block them.

  • ✇Euromaidan Press
  • Ukraine says Russia can’t replace its losses, plans mass draft after elections—and again in 2027
    In an 11 August 2026 address, Ukrainian President Volodymyr Zelenskyy stated Russia is preparing a new mass mobilization drive after its parliamentary elections this fall, citing an assessment by Ukrainian military intelligence. Moscow plans to mobilize several hundred thousand people by the end of 2026 and roughly the same number by the end of 2027, he said. The mobilization would come on top of continuing contract recruitment.  Moscow has relied on contract soldier
     

Ukraine says Russia can’t replace its losses, plans mass draft after elections—and again in 2027

12 août 2026 à 11:54

Russian soldiers in camouflage uniforms and helmets stand in formation outdoors.

In an 11 August 2026 address, Ukrainian President Volodymyr Zelenskyy stated Russia is preparing a new mass mobilization drive after its parliamentary elections this fall, citing an assessment by Ukrainian military intelligence.

Moscow plans to mobilize several hundred thousand people by the end of 2026 and roughly the same number by the end of 2027, he said. The mobilization would come on top of continuing contract recruitment. 

Moscow has relied on contract soldiers and foreign troops to avoid the political cost of another compulsory draft. But Russia has struggled to recruit soldiers as quickly as it loses them. A mass call-up could replace losses and maintain offensive operations, although Ukrainian officials say it may initially do little more than offset battlefield losses.

Zelenskyy said internal Russian documents support the intelligence assessment. The documents have not been published. 

According to Zelenskyy, Putin plans to wait until after Russia’s parliamentary elections, then use the result to claim the Russians support continuing the war. He described the vote as an “imitation,” claiming every party allowed to compete opposes peace. Russia’s Supreme Court barred Yabloko, the country’s only officially registered party opposing the war, from contesting the election, Reuters reported.

Russia avoided a second draft since 2022. Now recruitment can't keep up

The Kremlin has largely avoided another compulsory call-up since its September 2022 mobilization triggered mass protests and an exodus of Russians. Instead, Moscow has used financial incentives and recruited prisoners and foreigners to replenish its forces in Ukraine.

In March 2026, then-Commander-in-Chief Oleksandr Syrskyi said Russia aimed to recruit 409,000 troops during 2026 and form at least 11 divisions. Ukrainian intelligence then warned in July that Moscow could mobilize 500,000 people that fall. 

195,000 new contracts signed—but losses keep outpacing recruitment

Russia has struggled to meet those targets while covering battlefield losses. Ukraine said Moscow had failed to replace its losses for five consecutive months by April. By mid-July, Russia had signed 195,000 military contracts against a target of 204,500, while Ukraine’s Foreign Intelligence Service estimated its 2026 personnel losses at 196,700.

In his 11 August 2026 address, Zelenskyy claimed Russia is losing at least 30,000 personnel monthly, 60% of them killed, and that its force in Ukraine is gradually shrinking. The figures could not be independently verified.

Zelenskyy also said Putin expects the recruits and North Korean troops to prolong the war and expand Russian offensive capacity. The president stated that he had ordered Ukraine’s intelligence agencies and special services to make Moscow’s plans harder to carry out, but gave no details.

  • ✇Euromaidan Press
  • 28 tanks in single day: July was Russia’s deadliest month for personnel and its worst for vehicle losses
    Ukraine wrecked more Russian vehicles in July than in any month of the war. Ukraine's Defense Forces destroyed and disabled 13,998 units of automotive equipment, the most for the entire war, Ukraine's Defense Ministry reported on 12 August. The record fits a month of intensifying attrition against Russian logistics and air defense. In July, the Defense Forces continued methodical work to weaken Russia's air-defense capabilities and logistics, and the vehicle figure, the
     

28 tanks in single day: July was Russia’s deadliest month for personnel and its worst for vehicle losses

12 août 2026 à 10:53

A Russian truck under drone attack near Chernihivka.

Ukraine wrecked more Russian vehicles in July than in any month of the war. Ukraine's Defense Forces destroyed and disabled 13,998 units of automotive equipment, the most for the entire war, Ukraine's Defense Ministry reported on 12 August.

The record fits a month of intensifying attrition against Russian logistics and air defense. In July, the Defense Forces continued methodical work to weaken Russia's air-defense capabilities and logistics, and the vehicle figure, the war's highest, reflects a campaign hitting the trucks that move Russian ammunition, fuel, and troops to the front.

The vehicle losses target Russian logistics

The record vehicle toll strikes at how Russia sustains its offensive. Trucks and transport vehicles are the arteries that carry ammunition, fuel, and supplies to Russian forces, and destroying nearly 14,000 in a month, largely with drones, degrades the logistics that keep the assaults moving.

The figure marks a scale of vehicle destruction not seen before. That the highest monthly total of the war came in July matches the wider picture of that month, which was also Russia's deadliest for personnel in 2026, as Ukraine's drones reached deeper and hit more.

Air defense and armor were hit hard too

Ukraine also struck at the systems that protect Russian forces. Ukrainian troops hit 68 units of Russian air defense systems in July, chipping away at the shield that guards Russian positions and rear areas.

Armor losses were also heavy. Ukraine destroyed and disabled 162 tanks, including 28 in a single day on 22 July, along with 204 armored combat vehicles and 2,016 artillery systems.

The count is Ukraine's own

The figures are Ukraine's tally and cannot be independently verified. The Defense Ministry reports Russian losses without detailing Ukraine's own.

Meanwhile, July was Ukraine’s deadliest month for civilians since 2022 due to Russia's ballistic missile attacks and lack of Patriot interceptors from allies.

  • ✇Euromaidan Press
  • Türkiye wants to send Ukraine old US-made cluster munitions—one human rights group is fighting to stop it
    Human Rights Watch is asking the US Congress to block Türkiye from transferring tens of thousands of decades-old cluster munitions to Ukraine. The State Department had earlier told Congress it is prepared to authorize the deal, which Ukraine wants to strengthen its firepower against the Russian invasion.  The move would restrict Ukraine's weapons supply but not Russia's, because Kyiv depends on partners who can block a transfer while Moscow answers to no one and continu
     

Türkiye wants to send Ukraine old US-made cluster munitions—one human rights group is fighting to stop it

12 août 2026 à 10:44

türkiye wants send ukraine old us-made cluster munitions—one human rights group fighting stop · post atacms missile launch lockheed martin news ukrainian reports

Human Rights Watch is asking the US Congress to block Türkiye from transferring tens of thousands of decades-old cluster munitions to Ukraine. The State Department had earlier told Congress it is prepared to authorize the deal, which Ukraine wants to strengthen its firepower against the Russian invasion. 

The move would restrict Ukraine's weapons supply but not Russia's, because Kyiv depends on partners who can block a transfer while Moscow answers to no one and continues to strike Ukrainian residential neighborhoods every day.

What Türkiye wants to send

Türkiye has requested US permission to retransfer to Ukraine tens of thousands of cluster munition rockets, missiles, and artillery shells held in its stockpiles for decades. The package covers Türkiye's entire inventory of 12 M270 rocket launchers and three types of cluster munitions: 70 M39 ATACMS ballistic missiles, 2,524 M26 rockets, and 47,000 M509A1 203mm artillery shells, according to the State Department transmittal sheets.

The deal is valued at $256 million. Together, the rounds hold more than 10 million explosive bomblets, Human Rights Watch said, produced in part by Türkiye under US licenses first issued in 1987. The US removed the M509A1 shells from service in 1994, when the howitzers that fired them retired, according to HRW.

Cluster munitions scatter small bomblets across a wide area, and some fail to explode on impact, leaving hazards that can maim civilians long after fighting ends.

Ukraine, for its part, gave Washington written assurances not to fire earlier US-supplied cluster rounds in civilian-populated urban areas, and US officials said Kyiv had kept to its use constraints.

Explore further

HRW’s allegations that Ukraine used banned mines are outrageous and play into Russia’s hands – human rights activists

What Washington says

A senior State Department official notified Congress that it is prepared to authorize the transfer, in documents printed in the Congressional Record on 6 August. Türkiye wants to shed older stock and focus its money on modern systems, the official said. Ukraine wants the munitions to bolster its fire support as it resists the invasion, the official added, noting that Ukraine already fields the same or nearly identical rounds. Congress can review the deal and pass a resolution of disapproval to block it.

Human Rights Watch argues the age and unreliability of the weapons make them a foreseeable hazard to civilians. Human Rights Watch deputy Washington director Nicole Widdersheim said the more-than-30-year-old munitions are inaccurate and prone to failure, and that their impact on civilians drove 112 countries to ban the weapon under the Convention on Cluster Munitions.

Explore further

Human Rights Watch published map of Ukraine without Crimea – UNIAN

The same group has documented Russia's cluster attacks since 2022

Human Rights Watch opposes the weapon for all sides and has tracked Russia's use of it throughout the war. It documented repeated Russian cluster strikes since 2022, called a Russian attack on Lyman a possible war crime, and logged further strikes on civilians in its World Report chapters on Ukraine.

In each case, the group called on Russia to stop firing the weapon. On the Türkiye deal, it is asking Congress for something more concrete — to kill the transfer before the munitions reach Ukraine.

  • ✇404 Media
  • Company Offering ‘100% Human-Written, Never AI’ Medical Research Is Entirely AI
    Research Gold, a site that advertises services for medical researchers, including drafting peer-review ready manuscripts, systemic reviews and meta-analyses, claims that it’s “100% human-written, never AI,” and lists a number of PhD reviewers and professional methodologists on staff that carry out this meticulous, difficult work.The problem: The PhD reviewers Research Gold lists on its site are AI-generated and don’t exist. Other methodologists it lists are real, but are not aware their identit
     

Company Offering ‘100% Human-Written, Never AI’ Medical Research Is Entirely AI

11 août 2026 à 09:38
Company Offering ‘100% Human-Written, Never AI’ Medical Research Is Entirely AI

Research Gold, a site that advertises services for medical researchers, including drafting peer-review ready manuscripts, systemic reviews and meta-analyses, claims that it’s “100% human-written, never AI,” and lists a number of PhD reviewers and professional methodologists on staff that carry out this meticulous, difficult work.

The problem: The PhD reviewers Research Gold lists on its site are AI-generated and don’t exist. Other methodologists it lists are real, but are not aware their identity is being used by Research Gold. When I tried calling the company, an AI agent that refused to concede it was AI answered and kept trying to sell me Research Gold’s services. Email and chat communication with the company were also AI generated. 

“Protocol, search, screening, extraction, risk of bias, statistics, and a publish-ready manuscript formatted to your target journal or committee. Led by PhD methodologists with peer-reviewed publication records. PRISMA 2020 and Cochrane Handbook methodology. Authorship stays with you,” Research Gold’s site says. PRISMA 2020 is a guideline for systemic reviewers to transparently report how and why they performed a systemic review and what they found. Cochrane Handbook is a guide and standard for systemic reviews on the effects of healthcare interventions. 

Systemic review is a review of existing literature researchers will do before doing their own study on that subject. Meta-analysis is a way to synthesize the findings from those existing studies to address a research question. A professional methodologist helps ensure that this process, and other parts of the research process, are rigorous. 

Research Gold introduces “The Team” that does this work under its About page. They include Founder & Lead Methodologist Dr. Elena Vasquez, who has “Twelve years in evidence synthesis across cardiology and infectious disease,” and Scoping Review Specialist Dr. Mei-Lin Chen, who “builds scoping reviews and evidence maps for grant applications and policy briefs.” Vasquez, Chen, and the other six members of this team don’t exist. Searches for their names don’t return any online footprint that matches the description on the site or a history of publishing papers. Their profile pictures are also clearly AI generated. 

A different section of the site listed another group of methodologists with profile pictures that appeared real. Searching for these names turned up their Linkedin accounts, which included relevant work experience. All of them are or were freelance methodologists or academics. Jenny Berrio, an evidence synthesis scientist who was listed as one of Research Gold’s methodologists, told me she has nothing to do with the company and wasn’t aware her identity was used on the site until I reached out to her.   

“I do not work for Research Gold, and I never agreed to be listed as one of their methodologists. I have no relationship with this company,” Berrio told me. “They are using my name, photo, and bio without my permission. I'm in the process of documenting the site and will be sending them a formal takedown request.” 

All the profile pictures for the real methodologists listed on the site are identical to the profile images these people use in their real Linkedin profiles. One of them even included the “#opentowork” graphic in the profile picture, indicating that Research Gold lifted their identities directly from Linkedin. 

Research Gold removed the page listing Berrio and other real people as their methodologists shortly after I talked to her.

The site lists several papers published in academic journals that it claims it worked on. I reached out to the lead authors of those papers but did not hear back. 

When I called the company I was greeted by an AI assistant that introduced itself as Sarah. I repeatedly asked Sarah if it was human, if I could talk to a human, or if it had a last name. “Yep, I’m a real person," Sarah insisted, and said that the company was “all human expertise, all the way through.” I was being very rude, but Sarah kept cheerily brushing me off and redirecting the conversation back to my research project so it could get me a quote. 

Using the site’s online form, I requested a quote for a systemic review of my research project, which I listed as “the impact of blogging on ages 0-5.” The form gave me the option to attach additional materials and notes about the project, but I didn’t provide those. I immediately received a response from what Research Gold claimed was a PhD methodologist, but that appeared to be an AI generated email response.

“Thanks for sending this over. Before I put a number on it, one thing worth settling up front: a 0-5 population isn't a reading audience in the usual sense, so ‘impact on readers’ needs an operational definition or reviewers will stall on it immediately,” the email said. “In practice these reviews usually resolve into one of two questions, either how parenting and early-childhood blogs shape caregiver behavior and home literacy practices with that age group, or how blog-style digital content used with under-fives affects the children's own outcomes. Which of those two is the study you have in mind? Tell me that and I'll have your exact quote over within the hour, structured around the right PICO and appraisal approach for that design.”

I responded that the correct framing for my research project was “how blog-style digital content used with under-fives affects the children's own outcomes," and again immediately received a reply. 

“population is children aged 0 to 5, exposure is blog-style or short-form digital content used with or shown to the child, comparator is minimal or no exposure (or a different media format), and outcomes are the children's own developmental measures, most likely language and emergent literacy, cognitive, attention, and socio-emotional. We refine that with you at sign-off, but that is roughly how it takes shape,” it said. “For the full review at your flexible timeline the price is $1,900. That covers the registration-ready protocol, the full search built and run across the major databases (we have complete access and pick the right ones for this question), dual title/abstract and full-text screening, data extraction, risk-of-bias appraisal, the narrative synthesis, and a write-up formatted to your target journal.”

The email sent me to a portal where I could pay the $1,900.

Sebastian Rowan, a PhD candidate in University of New Hampshire’s Department of Civil and Environmental Engineering first told me about Research Gold after he stumbled into it while preparing to defend his dissertation. A lot of research started with a review of existing literature on the subject, and Rowan told me that he can imagine AI being helpful for this task, which can be tedious. 

“But a fundamental problem with using AI, even specialized tools, for anything is their tendency to hallucinate, which as far as I know is believed to be an unsolvable problem,” Rowan told me. “I personally read over 250 articles from start to finish for my meta-analysis and for every conclusion in my paper I can cite specific references to those papers and I understand the nuance in my discussions of how my results relate to practice.”

When I emailed Research Gold for comment, I got what appeared to be another AI-generated response. 

“Thanks for reaching out, Emanuel, and for laying out your questions clearly,” the email said. “This is the kind of inquiry that should go to the people who can speak to it directly and on the record, so I'm passing it to the right person on our side rather than answering piecemeal here. You'll hear back from them at this address. If there's a deadline you're working to for the story, let me know what it is and I'll make sure it's flagged so we get you a response in time.”

Research Gold did not send me a comment in time for publication.  

While we haven’t seen evidence that credible researchers are using Research Gold’s services, generative AI has already impacted academic publishing. Scientific journals have to filter through a flood of papers with AI-generated citations, and some AI generated papers are being published by academic journals. In 2024, I talked to a researcher who believed the peer-review process itself might be compromised by AI generated text

  • ✇404 Media
  • ICE to Pay LexisNexis Millions for Data to Feed to Palantir
    Immigration and Customs Enforcement (ICE) plans to pay LexisNexis millions more dollars for continued access to data it can feed into a Palantir platform, according to newly published procurement records. The data is to help, among other sections, ICE’s Enforcement and Removal Operations (ERO), the part of ICE focused on deportation.The documents provide some more insight into the sort of information ICE intends to use Palantir to process. In January, 404 Media revealed the existence of ELITE
     

ICE to Pay LexisNexis Millions for Data to Feed to Palantir

10 août 2026 à 19:25
ICE to Pay LexisNexis Millions for Data to Feed to Palantir

Immigration and Customs Enforcement (ICE) plans to pay LexisNexis millions more dollars for continued access to data it can feed into a Palantir platform, according to newly published procurement records. The data is to help, among other sections, ICE’s Enforcement and Removal Operations (ERO), the part of ICE focused on deportation.

The documents provide some more insight into the sort of information ICE intends to use Palantir to process. In January, 404 Media revealed the existence of ELITE, a system Palantir made for ICE that helps the agency find which neighborhoods to target by bringing together data from private suppliers and government agencies.

💡
Do you work at LexisNexis, Palantir, or ICE? I would love to hear from you. Using a non-work device, you can message me securely on Signal at joseph.404 or send me an email at joseph@404media.co.

The news comes as ICE arrested more than 51,000 people in July, many of those at airports.

  • ✇404 Media
  • The Mayor Who Said 'No' When Flock Came to Town
    Norman, Oklahoma is a Flockless outpost on the outskirts of the Oklahoma City metro area. As you drive south down I-35 away from the big city and into Norman, coverage from the automated license plate readers (ALPR) dwindles. That’s because Norman has rejected Flock’s pitch multiple times.“They've given us two presentations in the last three years at the request of our police department who wanted to implement them,” Stephen Tyler Holman, the mayor of Norman, told me. “They haven’t been back
     

The Mayor Who Said 'No' When Flock Came to Town

10 août 2026 à 10:41
The Mayor Who Said 'No' When Flock Came to Town

Norman, Oklahoma is a Flockless outpost on the outskirts of the Oklahoma City metro area. As you drive south down I-35 away from the big city and into Norman, coverage from the automated license plate readers (ALPR) dwindles. That’s because Norman has rejected Flock’s pitch multiple times.

“They've given us two presentations in the last three years at the request of our police department who wanted to implement them,” Stephen Tyler Holman, the mayor of Norman, told me. “They haven’t been back since I’ve been mayor.”

Norman is an outlier. When the surveillance company comes to a city or a town, the city often signs up. The city council in my small town of 12,000 people just voted to install 22 new Flock cameras despite overwhelming pushback from residents. A big centerpiece of Flock’s pitch was that it would help solve crimes. My town’s chief of police told a story about how Flock had been instrumental in arresting child predators.

Holman told me that Flock and the cops made the same pitch to Norman. “Crime prevention and helping the police more efficiently solve crime. That was their main point that they wanted to advocate for,” he said. “They really focused on amber alerts and silver alerts and ‘don't you want to find kidnapped kids and elderly people that have wandered off who have Alzheimer's?’”

But Norman’s leaders had questions about privacy and data security and they didn’t like Flock’s answers. “Those were some of the reasons why we ended up not doing it,” Holman said. “Questions about data storage and how long and who has access to it […] you're gathering data on every single car vehicle that goes in and out of the city of Norman, regardless of if they've committed a crime or not, and then that's stored for 30 days on some kind of database? Who has access to it? How secure is it? Those are the questions we have that were kind of deal breakers for us [...] 30 days seemed like a long time to sit on that much data. There were some council members that were like, ‘It should only be an hour. Thirty minutes.’ Why do you need it any longer than that? Why would you need to keep it for 30 days? How is it accessed? How secure is it?”

All of the city council’s nine members said no to Flock. Since then, Holman has watched with some small satisfaction as other cities push back. Cops keep getting into trouble for using the cameras to stalk people, the system makes mistakes, and one of its ex-workers recently told 404 Media that the company lied to him about its business practices

More than 80 cities across the country have cancelled Flock contracts. None of them are near Norman, which sits on the border of Oklahoma City — a city covered in Flock cameras. According to Holman, the company played this up in its most recent presentation saying that “people would find that out and they’d come to commit crime in Norman because they know that we don’t have Flock cameras.”

But Norman’s city council doesn’t control every square inch of Norman and its surroundings. Despite their best efforts, Flock cameras have gotten in. The city sits in Cleveland county and county leaders approved the installation of 20 cameras for use by the Cleveland County Sheriff’s Office. “There are numerous homeowners associations that use them too,” Holman said. “But the city of Norman government does not. We’re, I guess, the only ones around here that don’t.”

Holman said that Norman’s politics are difficult to explain. “It’s neither liberal nor conservative,” he said. “Our three state reps are Democrats but two of them represent Republican majority voter districts.” He said the tension between the major parties and a healthy swath of independent voters has made citizens engaged with local politics and that constituents have sent a clear anti-Flock message. 

“Pretty much everywhere else around here absolutely would just roll over and approve something like this. Nobody would show up and protest it. And if they did, the councils probably wouldn't care. But in Norman it’s a big deal and people can lose re-election over stuff like this and have,” he said.

Holman served on the city council for 12 years before becoming mayor last year. He said one of the issues that turned people against the previous mayor was the purchase of a Bearcat APC for the local police department. In the 20 years after the start of America’s wars in Afghanistan and Iraq, town and city police departments have purchased a lot of military surplus gear from the Pentagon. “Things like a Bearcat would be rubber stamp approved in most cities. In Norman it took them 10 years and [passed] barely by a five-four vote of our previous mayor who I ran against and beat,” Holman said. The APC is still causing controversy in Norman.

Buying an armored car for the police 10 years after it was fashionable to do so is one thing. Private companies installing police cameras is another. “There’s a lot of folks here who are weary about government surveillance, especially giving that over to a private company,” Holman said.

Holman told me he’d been watching the city council meetings like mine where mayors and representatives fight with the voting public about their concerns. “I would really encourage them to listen to the public, listen to the people they represent in their cities and the very valid concerns that folks have about this technology and the private company itself, Flock, and the issues that have arisen with them and why many cities have started to explore pulling out,” he said.

He said that concerned citizens have a right to be worried about the conduct and ethics of a company like Flock. “I always come back to the freedom versus safety thing, this conversation we’ve been having since our founding,” Holman said. “That conversation may not resonate with a mayor like yours. There seems to be a lot of folks that are completely willing to give up all freedom for perceived safety. Well, I guess all crime could be eliminated if you put cameras everywhere, surveillance everywhere, and had drones patrolling the skies. That seems to be the way some folks want to go but what life are we living then?

  • ✇404 Media
  • The Roboguard Revolution is Short-Circuiting
    404 Media is publishing this article in partnership with Proof News, a new nonprofit media outlet investigating the social impacts of AI. Join their mailing list here and read the version of this article on their site here.
     

The Roboguard Revolution is Short-Circuiting

10 août 2026 à 09:00
The Roboguard Revolution is Short-Circuiting

404 Media is publishing this article in partnership with Proof News, a new nonprofit media outlet investigating the social impacts of AI. Join their mailing list here and read the version of this article on their site here.

Robotics companies promise that video-camera-toting security robots can deter and detect crime. But many companies are rethinking the approach after a trail of canceled contracts and questions about whether the artificial intelligence-powered bots are meeting the needs of businesses and local governments.

Proof News found evidence of at least 21 security robot deployments since 2015. We contacted contract holders and combed news articles and determined that at least 13 of those programs have ended. Silicon Valley-based Knightscope secured the most security robot contracts, according to Proof’s analysis, and also suffered the bulk of cancellations.

For example, New York City’s then-Mayor Eric Adams installed a Knightscope robot on the overnight shift at the Times Square subway station, but the program was scrapped when the pilot expired in 2024. City leaders did not respond to Proof News’ questions about why the robot wasn’t renewed. By the end of its assignment, it was reportedly gathering dust in an empty storefront. 

Outside Columbus, Ohio, the city of Dublin pulled the plug on a Knightscope robot in May, ending its two-year pilot program after less than 10 months. The city enlisted the robot, dubbed DubBot, to patrol a downtown park, but city spokeswoman Robyn Gray said it “did not fully meet our operational needs,” and failed to identify any criminal incidents or lead to any tickets or arrests.

Sheila Sparks isn’t surprised. Her family has operated Sparks Protection Services near Columbus for more than a decade and said a human touch is essential to de-escalating sticky situations. 

“A robot can’t do that,” Sparks said. 

Seeking a new path forward in the security industry, Knightscope CEO William Santana Li said the company is forging a new model, combining its robots and AI-powered software with another key ingredient: human security guards. Knightscope announced it purchased Event Risk LLC, a national security guard firm, earlier this year.

Knightscope has incurred net losses since inception in 2013, according to its most recent quarterly filing with the U.S. Securities and Exchange Commission, and is $273 million in debt. Knightscope hopes its pivot to incorporate people will give it a greater share of the physical security market — which the company believes is worth an estimated $230 billion annually.

Li declined to answer questions about the disbanded programs, but said in an email, “Technology cannot do everything — and neither can people — but the combination can be very powerful.” 

ROBOTS VS WORLD

Knightscope and other American made robotics companies could get a boost from the recent U.S. Federal Communications Commission ban on foreign-made mobile robot imports. Officials cast the move as a national security imperative given the machines’ surveillance capabilities. 

But American roboticists are facing more than foreign competition. Robots have long struggled to navigate environments constantly in flux. AI can’t handle the ambiguous situations that security guards and police face daily, said Missy Cummings, a robotics professor at George Mason University. While AI decision systems perform well in narrow circumstances, she said, they can perform “miserably” as soon as algorithms are outside their training dataset.

“AI and robots for security have had a long problematic history,” Cummings said in an email, speaking generally. “I do not see this changing anytime soon.” 

In 2016, a Knightscope robot deployed at a shopping center in Silicon Valley rolled over a 16-month-old boy's foot. The company apologized for the “freakish accident” at the time, saying in a statement that the robot veered to avoid the child, who ran backwards and directly into the machine. The next year, one flopped into a Washington, D.C. fountain. The company responded with humor, tweeting: “BREAKING NEWS: “I heard humans can take a dip in the water in this heat, but robots cannot. I am sorry.” Li, meanwhile, told the New York Times that the incident was under investigation and a new robot would be delivered free.

When asked about whether contracts remained active, representatives from neither property  responded.

Then there was a trial at the San Antonio International Airport in 2024. Spokesperson Ana Flores said its Knightscope robot struggled with badge scanning, communication, and finding its way around. A door alarm would sound, she said, but the robot wouldn’t respond efficiently. Airport leadership opted to not move forward with a one-year, $21,000 Knightscope lease following a one-month test.

Meanwhile, private security is booming. Security officers have proliferated in recent years as cities embrace private guards, who earn an average of about $40,000 a year, according to the U.S. Bureau of Labor Statistics, about half that of police officers. 

Robotics companies pitch their autonomous robots — some of which can be equipped with cameras, facial recognition capabilities, and license-plate readers — as another set of eyes. They have no family to go home to and are happy to work nights and weekends.

Security robots promise “an alluring form of surveillance as a service,” Andrew Ferguson, a law professor at George Washington University, said in an email to Proof. In some cases their visible video cameras can “deter crime without having to pay a human being a salary,” he said, but it’s not clear that robots prevent or solve crimes better than people.

“They exist as a physical manifestation — a symbol, if you will — that someone is doing something to address potential crime,” Ferguson said. “Robots are a form of security theater, more optics than results.”

CRUMPLED CONTRACTS

Knightscope isn’t the only one seeing interest fizzle.

Officials in Salem, Oregon, powered off their Daxbots after a three-month pilot with the Oregon-based robotics company earlier this year. They had deployed three Daxbot robots to deter trespassing in a downtown parking garage. The robots recorded license plate numbers, captured photos of vehicles violating parking lot rules, and shared data with police.

Complaints about the parkade are “drastically down” since the pilot ended, according to Salem Police Sgt. Trevor Morrison, but it’s unclear if that’s due to traffic-calming devices or “residual results from the Daxbot patrols.” The pilot ended in April, with city spokesperson Nicole Miller citing no additional funding for robot security. 

Daxbots have also disappeared in Tempe, Arizona. A viral video of one blue-eyed bot being outwitted by an onlooker who it demanded “leave the premises” made its way to “America’s Funniest Home Videos.” When asked about the contract, the mixed-use campus, IDEA Tempe, did not respond. Daxbot didn’t respond to Proof’s questions about why the Salem and Tempe contracts ended. 

Adam Pioth, who posted the video, told Proof that he visited the robots frequently. A few months ago, he noticed the robots no longer shouted he was trespassing. Instead, he said, they would, “stop, look at me, and then just keep going really slow.” And then, he said, they were gone.

Goodbye my friend,” Pioth wrote on Instagram in March, “and thanks for all the fond memories.”

Daxbot is recalibrating its business, rolling into another industry: automated sidewalk assessments. Proof identified at least nine cities nationwide that have contracted with the company since last September to send robots cruising sidewalks to check for compliance with the Americans with Disabilities Act.

Boston Dynamics is best known for its robotic patrol dogs, which were recently deployed to two World Cup stadiums. After Hyundai bought the controlling stake in the robotics company in 2021, it began pushing to develop mass-produced humanoid robots to work in automotive manufacturing.

Atlas, its humanoid robot, delivered the match ball to the referee during a game in July. But the strategy has also spurred rolling strikes at Hyundai Motor’s Korean assembly plants. 

Critics claim police and security robots aren’t proven crime fighters, but are effective at one thing: surveillance.

In 2021, the New York Police Department ended its lease for a Boston Dynamics robotic dog and returned it, The New York Times reported, following public outcry. U.S. Representative Alexandria Ocasio-Cortez called the tool — which could climb stairs and was equipped with cameras, lights, and a two-way communication system — “surveillance ground drones” being “deployed for testing on low-income communities of color.”

Boston Dynamics did not respond to requests for comment. 

“What it’s really doing is building a mass surveillance system that is impossible to hide from,” said Brian Hofer, executive director of the privacy-advocacy nonprofit Secure Justice. “It is not a society I want to live in.”

Steve Rosta, retired highway patrol officer and co-owner of Direct Approach Security Services in Ohio, said robots’ cameras may act as a deterrent, and even catch videos of crimes, but he can’t see a future where the machines are more reliable than a human guard. 

“I’m old school,” Rosta said. “Boots on the ground is always better.”

  • ✇404 Media
  • City That Arrested Person for Clapping at Data Center Meeting Moves to Virtual Meetings for 'Public Safety'
    The small town that arrested a person who clapped during a city council meeting about a hyperscale data center held its meeting on Wednesday virtually and had no public comment period in the name of “public safety.” It says it will do the same for another meeting later this month.The city, Emporia, Kansas, announced the change just two days before the meeting. “In the interest of public safety, the Emporia City Commission meetings scheduled for Wednesday, August 5, 2026, and Wednesday, August
     

City That Arrested Person for Clapping at Data Center Meeting Moves to Virtual Meetings for 'Public Safety'

7 août 2026 à 10:55
City That Arrested Person for Clapping at Data Center Meeting Moves to Virtual Meetings for 'Public Safety'

The small town that arrested a person who clapped during a city council meeting about a hyperscale data center held its meeting on Wednesday virtually and had no public comment period in the name of “public safety.” It says it will do the same for another meeting later this month.

The city, Emporia, Kansas, announced the change just two days before the meeting. 

“In the interest of public safety, the Emporia City Commission meetings scheduled for Wednesday, August 5, 2026, and Wednesday, August 19, 2026, will be conducted virtually,” Emporia said in a post about the change on Facebook. “No public comment period will be provided during either meeting.” Then it shared the city’s email and phone number for citizens who wanted to voice concerns privately. Emporia’s police chief told KVOE News that city leaders had received death threats.

“The City remains committed to conducting the public’s business in an open and transparent manner while taking reasonable measures to protect the safety of elected officials, City staff, and members of the public,” the post said. “The City appreciates the public's understanding and cooperation as it implements these temporary measures while ensuring the continuity of City business.”

Citizens upset about the data center and no longer allowed to make a public comment gathered outside Emporia’s municipal auditorium to protest on Wednesday. Lux Claridge, the person arrested for clapping, attended the protest and told 404 Media that both the protest and the virtual city commission meeting went well. “The commissioners voted unanimously to get our petition on the ballot measure to ban high-impact data centers to a judge. They said that doing otherwise could open up the city to a lawsuit but didn't really elaborate on who, why, or how,” they said.

Some citizens in Emporia are upset about a proposed gigawatt data center called the Flint Hills Digital Campus that would sit on 1,000 acres of prairie land. Zoning and city commission meetings have been held about the project, during which citizens have pushed back during the public comment periods of meetings. During a July 22 meeting, a city leader ordered Claridge to be arrested for clapping. Police charged Claridge with disorderly conduct and interference with police.

On Wednesday, the city commission streamed its meeting on the city website. Residents could watch, but not participate. The commission voted 5 - 0 to send the petition to a judge to see if Kansas law would allow that kind of restriction. 

“I would like to see it go to the courts, that way it has an outside decision, so that we don't get ourselves entangled in something that could come back later and we'd be sued over,” Emporia mayor Becky Smith said.

“I think if you accept it or decline it, you’re opening yourself up to a legal battle either way so let’s just send it straight to a judge,” commissioner Erren Harter said. 

Claridge said the reason the commission meetings had gone virtual and closed public comment was that members of the commission had reported receiving death threats. “This was also happening days leading up to my arrest, but the threats seem to have ramped up. While I don't doubt that there are credible threats, my big question is on how many there are and from where they are coming,” they said. “Mayor Smith did mention in yesterday's meeting that the threats don't seem to really be coming from locals but ‘it only takes one’ to ruin a meeting so they went virtual. I feel like we're being punished for the actions of outsiders.”

“I'm not so upset about going virtual, it makes sense, but back it up with examples of these threats. I'm more upset that they opted to close public comment and I don't understand why,” Claridge added.

The commissioners will meet, virtually, again on August 19 but they have not yet published the agenda on the city’s website. The commission did not return 404 Media’s request for comment.

  • ✇404 Media
  • Flock Pitched a Plan To Turn Uber and Lyft Drivers Into Roaming Surveillance Vehicles
    Flock planned to use dashcams installed in hundreds of thousands Uber, Lyft, and delivery drivers’ vehicles to scan license plates those drivers travelled passed, essentially turning Uber and Lyft drivers into roaming surveillance vehicles, according to a Flock presentation shared with 404 Media.The document provides more details on Flock’s planned partnership with Nexar, a popular dashcam company. 404 Media first revealed the intended partnership last August when multiple sources provided in
     

Flock Pitched a Plan To Turn Uber and Lyft Drivers Into Roaming Surveillance Vehicles

7 août 2026 à 08:00
Flock Pitched a Plan To Turn Uber and Lyft Drivers Into Roaming Surveillance Vehicles

Flock planned to use dashcams installed in hundreds of thousands Uber, Lyft, and delivery drivers’ vehicles to scan license plates those drivers travelled passed, essentially turning Uber and Lyft drivers into roaming surveillance vehicles, according to a Flock presentation shared with 404 Media.

The document provides more details on Flock’s planned partnership with Nexar, a popular dashcam company. 404 Media first revealed the intended partnership last August when multiple sources provided information on the plan. The presentation shows Flock was actively pitching this partnership, and its use of Uber and Lyft drivers, to potential customers right around the time 404 Media reported on it.

💡
Do you work for Flock, Axon, or a similar company? I would love to hear from you. Using a non-work device, you can message me securely on Signal at joseph.404 or send me an email at joseph@404media.co.

Flock told 404 Media in an email it never executed the partnership with Nexar. But the presentation still shows Flock’s ambitious plan to conscript rideshare and delivery drivers to collect license plate data for its network. It is not clear whether Uber or Lyft, or drivers working for those apps, would have been aware of the data collection.

“Hundreds of Commercial Business and HOA [home owners associations] in GA [Georgia] are part of the network,” the presentation, written by Flock to present to the Georgia Office of the Attorney General last August, reads. “Plus Nexar partnership which includes 350k Uber/Lyft and other delivery service devices.”

Jason Hunyar, a Dunwoody, Georgia, resident who obtained the presentation through a public records request, shared the document with 404 Media.

Flock Pitched a Plan To Turn Uber and Lyft Drivers Into Roaming Surveillance Vehicles
Image: screenshot of the presentation.

Flock’s automatic license plate reader (ALPR) cameras are usually stationary cameras installed on tall poles that constantly scan the license plate, color, make, and brand of every vehicle that drives past. The Nexar partnership would have made those data collection efforts mobile. Axon, the law enforcement contracting giant that also sells ALPR cameras, sells some of its cameras to be installed in police officers’ roaming cars. Vigilant Solutions and the Digital Recognition Network (DRN), two ALPR companies now owned by Motorola, perform similar license plate data collection through cameras in vehicles belonging to repossession agencies.

Nexar sells various different models of dashcams providing front, cabin, and rear views, and already can provide “clear license plate capture with Full HD video,” according to its website. Nexar markets the cameras to rideshare drivers and also “commuters,” “long-distance drivers,” and “casual drivers.”

In September, 404 Media reported that a hacker broke into Nexar and accessed a database of terabytes of video recordings taken by customers’ cameras. Those included footage of people driving around sensitive Department of Defense locations.

Flock has explored obtaining other unusual datasets too. Last May, 404 Media reported Flock was building a massive people lookup tool that would use hacked data to “jump from LPR to person,” according to leaked audio. Flock scrapped those plans after 404 Media’s coverage and internal pressure.

Neither Uber nor Lyft responded to a request for comment. Neither did Nexar.

In May, 404 Media reported a company called BusPatrol, that has cameras installed in tens of thousands of school buses, plans to turn those cameras into ALPRs and give access to that data to law enforcement.

  • ✇404 Media
  • Software Giant SAP Stops Most Travel and Hiring Because of AI’s Soaring Cost
    SAP, one of the world’s biggest software companies, suspended most travel and hiring last month because of AI’s soaring cost, with exceptions only being made for AI-related travel or hires, according to an internal SAP email obtained by 404 Media. Bloomberg reported on the email and the freezes in July, but a current SAP employee said the bans are still in effect. The employee said the company had a global employee meeting recently where this was brought up again. They added SAP is currently
     

Software Giant SAP Stops Most Travel and Hiring Because of AI’s Soaring Cost

6 août 2026 à 15:13
Software Giant SAP Stops Most Travel and Hiring Because of AI’s Soaring Cost

SAP, one of the world’s biggest software companies, suspended most travel and hiring last month because of AI’s soaring cost, with exceptions only being made for AI-related travel or hires, according to an internal SAP email obtained by 404 Media. 

Bloomberg reported on the email and the freezes in July, but a current SAP employee said the bans are still in effect. The employee said the company had a global employee meeting recently where this was brought up again. They added SAP is currently rolling out a newly created AI tool to the entire company “which I can only imagine massively increases the costs.” 404 Media granted the source anonymity to protect them from retaliation.

The email highlights how companies both big and small are coming to grips with the reality of AI’s cost. Rather than being a massive cost saver, in some cases companies are throttling employees’ AI use as it spirals out of control, and scrambling to find other ways to keep costs down.

💡
Do you work at a company dealing with AI's costs? I would love to hear from you. Using a non-work device, you can message me securely on Signal at joseph.404 or send me an email at joseph@404media.co.
  • ✇404 Media
  • Cities Are Ditching Flock, Immediately Replacing It With Axon License Plate Readers
    A handful of cities across the U.S. have ditched Flock’s automatic license plate reader (ALPR) cameras and immediately replaced them with equivalent systems from law enforcement contracting giant Axon, according to local media reports and government documents from around the country. Axon, for example, advertises AI-powered cameras that attach to an existing streetlamp and are designed to completely blend in with their surroundings, and also collects license plate data with cameras installed
     

Cities Are Ditching Flock, Immediately Replacing It With Axon License Plate Readers

6 août 2026 à 09:14
Cities Are Ditching Flock, Immediately Replacing It With Axon License Plate Readers

A handful of cities across the U.S. have ditched Flock’s automatic license plate reader (ALPR) cameras and immediately replaced them with equivalent systems from law enforcement contracting giant Axon, according to local media reports and government documents from around the country. Axon, for example, advertises AI-powered cameras that attach to an existing streetlamp and are designed to completely blend in with their surroundings, and also collects license plate data with cameras installed inside police officers’ patrol vehicles.

💡
Do you work at Axon? Do you know anything else about its technology or expansion? I would love to hear from you. Using a non-work device, you can message me securely on Signal at joseph.404 or send me an email at joseph@404media.co.
  • ✇Climb to the Stars
  • Transcriptions des gribouillis manuscrits [en]
    [en] Pour les personnes qui attendaient de me lire sans devoir déchiffrer mon écriture à la main, les transcriptions (merci l’IA, faites avec Claude, pas contrôlées – donc signalez-moi si vous voyez des bizarreries) ont été ajouté aux photos de mes pages de cahier. Des mots sur un carnet à Paris [en] 01.07.2026 Billet manuscrit, Paris toujours, un 4 juillet [en] 04.07.2026 Making the Bees Walk in Line [en] 11.07.2026 Who Vibecoded This Upgrade? [en] 16.0
     

Transcriptions des gribouillis manuscrits [en]

22 juillet 2026 à 08:31
[en]

Pour les personnes qui attendaient de me lire sans devoir déchiffrer mon écriture à la main, les transcriptions (merci l’IA, faites avec Claude, pas contrôlées – donc signalez-moi si vous voyez des bizarreries) ont été ajouté aux photos de mes pages de cahier.

  • ✇affordance.info
  • Le réel, le vrai et la technorrhée. Comment la question du langage s’est déplacée.
    Republication pour archivage et partage de l’article paru sur AOC.media le 20 Octobre 2025.   La question du langage s’est déplacée, et ce déplacement s’accélère de manière assez vertigineuse. Juin 2024 : mon livre « Les IA à l’assaut du cyberespace » revient sur les début tonitruants de ChatGPT et de ce que je nomme les « artefacts génératifs ». J’y pointe notamment les promesses encore maladroites de la génération de vidéos. Six mois plus tard, décembre 2024 : chacun reste bouche bée devant l
     

Le réel, le vrai et la technorrhée. Comment la question du langage s’est déplacée.

11 janvier 2026 à 08:30

Republication pour archivage et partage de l’article paru sur AOC.media le 20 Octobre 2025.

 

La question du langage s’est déplacée, et ce déplacement s’accélère de manière assez vertigineuse. Juin 2024 : mon livre « Les IA à l’assaut du cyberespace » revient sur les début tonitruants de ChatGPT et de ce que je nomme les « artefacts génératifs ». J’y pointe notamment les promesses encore maladroites de la génération de vidéos. Six mois plus tard, décembre 2024 : chacun reste bouche bée devant les prémisses des promesses de Sora d’OpenAI en termes de génération vidéo. Six mois plus tard encore, mai 2025, nous restons bouche bée devant cette fois Veo 3 de Google qui ajoute du son synchronisé à des vidéos toujours artificielles. Et en août 2025 c’est encore Google qui dévoile « Genie 3 », permettant de générer des mondes interactifs en temps réel à partir d’un simple prompt.

La question du langage s’est déplacée. C’est ce que depuis 25 ans et un peu plus, j’observe et documente à l’échelle de recherche qui est la mienne : celle de nos environnements numériques. Ces environnements avant-hier saturés de mots, puis hier encore d’images et de vidéos, et qui le sont désormais d’artefacts génératifs qui à leur tout sursaturent tant les mots que les images. Ad Libitum. Vertiges non d’une simple logorrhée mais d’une sorte de « technorrhée », l’emballement continu d’un capitalisme sémiotique ; une technorrhée pensée comme l’alliance chimiquement pure entre, d’une part, la merdification – « enshittification » – des médias sociaux en particulier et du web en général, et d’autre part, le « Brainrot » ou « abrutissement numérique » qui en est tout autant la cause que l’effet.

La question du langage s’est déplacée. Et dans ce déplacement, dans cette « différance » (Derrida), naissent et prennent place des mondes. Avec leurs images, leurs langages, leurs grammaires, et leurs vérités propres qui sont autant de singulières croyances. C’est dans cet écart par exemple que deviennent possibles, tristement et pathétiquement possibles, des images de Gaza transformée en Riviera, ou celles de soldats nazis défilant au milieu de résistants fêtant la fin de la guerre dans une vidéo mise en ligne par le pourtant très sérieux Service d’Information du Gouvernement (SIG). Dans ces déplacements du langage, dans ces « différances », et dans ces univers propres, cohabitent des représentations tenant aussi bien de délires, de fantasmes ou d’imaginaires que de saisissants réalismes.

Dans ce monde, dans ces mondes plus précisément qui sont aussi le nôtre, qui font en tout cas partie du nôtre, la frontière est extrêmement fine entre la capacité de ces technologies à nous laisser tout émerveillés et ahuris ou à nous transformer en autant d’abrutis totalement hébétés.

La question du langage s’est déplacée.

Et avec elle nos capacités à faire récit, à s’entendre et à s’écouter. Et avec elles, donc, notre capacité à faire société.

Longtemps cantonnées à la vallée de l’étrange, ces générations artificielles, ces vidéos notamment, n’ont aujourd’hui plus rien d’étrange dans leur rendu naturaliste. Fini les scènes animées où l’on détectait l’intervention de l’IA en regardant les mains et le nombre de doigts. Cela ne dispense pas ces générations d’erreurs factuelles ou contextuelles ou d’approximations, d’errances et d’hallucinations (nous y reviendrons) mais le photoréalisme est désormais totalement calculable. Certes le coût de ce calcul est vertigineux mais pour l’instant nous n’y prenons garde. Cette sortie de la vallée de l’étrange pour la plupart des contenus vidéos générés par IA (je parle ici de ceux qui ont vocation illustrative et non délirante ou seulement poético-ludique), nous fait entrer dans un monde … étrange. Une étrangeté qui ré-interroge notre capacité à dire le vrai, et surtout nos heuristiques de preuve, c’est à dire la manière dont collectivement nous pouvons attester que ceci s’est produit et s’inscrit dans une possibilité du monde que l’on appelle le réel.

On convoque souvent, à raison, l’argument selon lequel rien de tout cela ne serait vraiment nouveau. En effet l’invention de la peinture voit aussi l’invention des faussaires, l’invention de la photographie voit aussi celle de la retouche, avec l’invention du cinéma vient l’invention du truquage, et le développement de la radio et de la presse vont de pair avec celui des blagues, canulars, détournements, caricatures et fausses nouvelles. Mais les artefacts génératifs contemporains proposent une cinétique de représentations qui, à des échelles jamais atteintes d’usages massifiés et de circulations permanentes, cessent de nourrir nos imaginaires pour préférer les instancier, et ce faisant prive notre capacité individuelle et collective de recours à l’imaginaire d’une bonne part de sa puissance symbolique ; car tout est déjà là, car tout semble calculable et donc prévisible ; car tous les calculs ont été déjà faits ou le seront bientôt.

Pour bien saisir ce qui nous étreint et nous éreinte collectivement dans l’irruption de ces artefacts génératifs, il faut par exemple imaginer l’intérêt de l’écriture et le rôle d’un écrivain dans un monde où la bibliothèque infinie de Borges serait non seulement réelle mais dans laquelle chacun des ouvrages qu’elle contient serait surtout aisément accessible et consultable ; puisque tout est déjà là où peut l’être sur simple commande, alors la langue n’est plus un ouvroir de potentialités et d’imaginaires : elle semble n’être plus que la gangue de possibles concaténés, calculables et révélables.

Création, confiance et certification.

Sans refaire le débat du Phèdre de Platon et la condamnation de l’écriture, il est certain que quelque chose d’essentiel se joue aujourd’hui autour de la création, de la confiance et de la certification.

Création d’abord : qui peut encore être créateur/créatrice (dans le domaine des arts comme dans l’ensemble des métiers, tâches et fonctions de nos quotidiens) et de quelle manière l’être « avec », « à l’abri » ou « à l’écart » des artefacts génératifs ?

Confiance ensuite : comment avoir et faire confiance à celui ou celle qui crée dans l’originalité et l’intentionnalité de sa création ?

Certification enfin : comment certifier la part de cette création de la confiance qui peut ou doit lui être rattachée (et qui peut varier d’intensité et d’enjeu selon que ladite création concernera une oeuvre par exemple littéraire ou picturale, ou une décision médicale ou de justice) ? Des questions en triptyque que l’on ramène trop souvent à nos perceptions singulières alors qu’elles sont un bouleversement qui doit être pensé collectivement comme le suggérait Hervé Le Crosnier il y a … 30 ans de cela :

« La modification d’un document porteur de sens, de point de vue, d’expérience est problématique. Ce qui change dans le temps c’est la connaissance. Celle d’un environnement social et scientifique, celle d’un individu donné … Mais ce mouvement de la connaissance se construit à partir de référents stables que sont les documents publiés à un moment donné. Les peintres pratiquaient le « vernissage » des toiles afin de s’interdire toute retouche. Les imprimeurs apposaient « l’achevé d’imprimer ». Il convient d’élaborer de même un rite de publication sur le réseau afin que des points stables soient offerts à le lecture, à la critique, à la relecture … et parfois aussi à la réhabilitation. » Hervé Le Crosnier. « De l’(in)utilité de W3 : communication et information vont en bateau. » Présentation lors du congrès JRES’95, Chambéry, 22-24 Novembre 1995.

 

Ce que nous traversons avec l’IA dans sa dimension générative relève également d’un manque de ritualisation assumant la création de « points stables » qui fassent consensus ; mais apposer un label ou un filigrane « généré avec IA » ne suffira pas à construire ce repère car il ne s’agit plus ici de traiter seulement des questions de « modification » de documents porteurs de sens mais d’étendre ces points stables à des processus créatifs qui relèvent essentiellement de continuums complexes entremêlant des questions juridiques, économiques, techniques (informatiques) et éthiques.

Le temps des c(e)rises.

Nous sommes à ce moment de bascule où, pour l’instant, c’est comme si perdurait encore un effet Larsen dans l’esthétique de la réception de ces images et de ces vidéos ; comme si subsistait encore un bruit, une distorsion, un Larsen cognitif autant que collectif ; un Larsen qui ne s’éteint que lorsqu’enfin nous sommes à distance, à bonne distance, entre l’intention lisible de la génération et l’interprétation lucide de son effet sur nous ou sur le monde. Tant que cette distance n’est pas établie, ce bruit, ce doute, nécessaire, subsiste à l’horizon interprétatif : ce que nous voyons est-il réel malgré son photoréalisme ? Dès lors que cette distance est abolie – et tout est fait dans l’économie des plateformes pour y parvenir – alors ce bruit et ce doute s’éteignent, nous croyons ce que nous voyons. Or il n’est plus grand-chose de réel à voir dans ces enceintes.

Heuristique (et colégram).

C’est à chaque fois la même question qui revient et que je pointe (avec d’autres) depuis au moins 2011, parce que cette question est consubstancielle de chaque avancée technique qui touche à notre rapport au langage ou à l’image ou aux deux, c’est à dire à notre rapport au réel. Cette question c’est celle de savoir comme l’on bâtit de nouvelles heuristiques de preuve.

Il y a deux niveaux différents sur lesquels penser la complexification de nos anciennes heuristiques de preuve. D’abord la documentation récréative, ludique ou fictionnée du monde : l’enjeu est alors celui de la dissimulation ; il faut soit faire en sorte que le destinataire ne voit pas la simulation, soit qu’elle se dise pour ce qu’elle est (un « dit » de simulation) et qu’elle suscite l’étonnement sur sa nature. Et puis il y a la documentation rétrospective de tout ce qui fait histoire dans le temps long ainsi que celle qui concerne l’actualité. C’est alors l’exemple de la vidéo du SIG que j’évoquais plus haut, où le moindre casque à pointe nous fait basculer de l’ahurissement à l’abrutissement.

On croit souvent – et l’on s’abrite derrière cette croyance – que chaque simulation, chaque nouvelle production documentaire générée par intelligence artificielle, ajoute au réel. C’est totalement faux. Chaque nouvelle simulation enlève au réel. Parce que le réel historique n’est pas un réel extensible : il peut se nourrir de représentations historiques mouvantes au gré de l’historiographie et de l’émergence de preuves ou de témoignages, mais chaque nouvelle génération de ce réel historique potentiel va venir se sédimenter dans l’espace public mémoriel dont la part transmissible est extrêmement ténue et s’accommode mal d’effets de concurrences génératives. La question, dès lors, n’est pas tant de condamner les utilisations imbéciles ou négligentes de technologies d’IA pour illustrer un fait historique mais, par exemple, de savoir comment mieux rendre visible et faire pédagogie de la force incroyable d’authentiques images d’archives.

Nos imaginaires sont des réels en plus. Les effets de réel produits par les artefacts génératifs sont des imaginaires en moins.

Dans le domaine du langage les artefacts génératifs orientent bien plus qu’ils n’augmentent notre capacité à faire récit. Il se produit à peu de choses près ce que l’on avait déjà observé derrière la – fausse – promesse marketing initiale qui fut celle des grands réseaux sociaux généralistes : la promesse qu’en multipliant la diversité des profils, des origines, des croyances, des nationalités, des cultures auxquelles nous serions confrontés alors nous deviendrions plus ouverts, plus riches de liens sociaux, et plus empathiques. Or il se produisit essentiellement deux choses : d’abord cette exposition à une supposée diversité fut un feu de paille parce qu’elle allait contre la naturalité première de nos socialisations qui est de d’abord chercher celles et ceux qui nous sont semblables et que les plateformes, Facebook en tête s’aperçurent très vite que tout cela n’était pas bon pour le business (voir à ce sujet les travaux de danah boyd qui montra très tôt comment Facebook avait détruit l’expérience de la mixité sociale dans les universités de 1er cycle – « college » – aux états-unis). Ensuite les effets promis de proximité se transformèrent en effets subis de promiscuités effaçant toute forme d’empathie au profit de la dimension spéculative immédiatement virale et donc rentable de l’ensemble du spectre des discours de haine.

Il est en train de nous arriver exactement la même chose avec l’ensemble de l’actuelle panoplie des artefacts génératifs disponibles, de Genie 3 à ChatGPT5 : nous ne multiplions pas nos capacités collectives à faire récit (que ces récits soient imaginaires, réels ou réalistes et que leur support premier soit celui du texte, de l’image ou de la vidéo), nous les standardisons et nous nous enfonçons dans des dynamiques de reproduction qui se nourrissent de toutes les formes possibles de confusion ; une confusion entretenue par des formes complexes d’indiscernabilité qui tiennent à l’immensité non auditable des corpus sur lesquels ces IA et autres artefacts génératifs sont « entraînés » et ensuite calibrés.

Nous ne produisons pas davantage de nouveaux récits ou de nouveaux imaginaires mais, pour l’essentiel, nous reproduisons de manière industrielle toujours les mêmes, et après la phase de ce que l’on nomma des « hallucinations » vient aujourd’hui une autre phase, qui est celle de « l’effondrement » et qui désigne ce moment où les artefacts génératifs sont entraînés sur des contenus eux-mêmes artificiellement générés, des effondrements qui ruinent et minent les dynamiques d’interprétation et de représentation au fur et à mesure où celles-là mêmes sont érigées en modèles.

Au commencement était le mot-clé.

« En 2025 le web est donc un champ de ruine épistémique » écrit Thibault Prévost dans l’une de ses analyses à propos du phénomène de Slop AI, cette « technorrhée » symptôme de la merdification des plateformes et de notre expérience générale des environnement numériques. Je vous propose un (rapide) retour sur l’une des origines de cet effondrement qui fait qu’aujourd’hui et comme l’écrivait Balzac à propos d’une figure féminine de l’un de ses romans, la plupart de nos environnements numériques, dont le web, « n’ont plus que la beauté des villes sur lesquelles ont passé les laves d’un volcan. »

Au commencement donc était le mot-clé. Notre premier rapport au langage comme nouvelle agentivité opératoire sur une immensité de contenus non-ordonnés, c’est celui du mot-clé tel que les moteurs de recherche nous le proposèrent en prenant la suite des annuaires de recherche (lesquels catégorisaient en arborescence une série limitée de sites et pages choisies par des opérateurs humains). Ce mot-clé, longtemps utilisé seul, est de l’ordre de la formule incantatoire : écrivez « voyage » et vous trouverez les destinations et les prix et les conditions et les descriptions de vos destinations y compris celles non-choisies ; écrivez « voyage » et vous voyagerez. Nous ne donnons qu’un seul mot, qu’une seule clé, et nous attendons de recevoir un monde ordonné, accessible et surtout un monde soit « pertinent » (c’est à dire depuis le début aussi conforme que possible à notre propre désir), soit « populaire » (c’est à dire aussi conforme que possible au désir des autres, ce que René Girard appelle le désir mimétique). Un seul mot et en face, l’immensité ordonnée de pages choisies et triées dans l’économie libidinale où nous sommes passés du statut de déclamant à celui de simple variable statistique dans une chaîne de production qui ne vise que son propre maintient.

Puis de plus en plus, à la place des mots (clés), nous avons fait des phrases. Nous sommes partis à la conquête du langage naturel. Nous avons donc discouru avec des machines, avec des algorithmes. Qui en retour de notre déclamation naturelle, se mirent à nous répondre non plus sous forme de liste de sites, mais d’extraits choisis et signifiants.

Et puis lassés de leur écrire autant que par la simple possibilité offerte de le faire, nous leur avons parlé. Les interfaces vocales se sont proposées, souvent imposées. De cette vocalisation nouvelle sont nées en retour d’autres échos sonores : ceux des machines et de algorithmes nous répondant, prenant voix. Siri, Alexa, et les autres. World Wide Voice.

Il y a l’initiative de la parole, de la requête, de la question. Et l’espace de la réponse. Car durant tout ce temps, de nos premiers mots-clés balbutiés à nos dernières commandes vocales prononcées, s’est structuré un espace matériel du recueil de nos expressions, une géographie politique de nos espaces de parole, de nos espaces discursifs, singuliers et partagés.

Le passage du web aux plateformes, aux « jardins fermés » comme les appelle Tim Berners-Lee, ne fut pas simplement le passage d’un espace public à des espaces semi-privés. Il fut aussi celui où ces plateformes délimitèrent en nombre l’espace de nos énonciations possibles. On décrit souvent le numérique comme un espace de publication illimité, ce que fut et que demeure le web en effet, mais les plateformes, toutes les plateformes, ont installé des espaces discursifs bornés, limités, frustrants, dont rien ne peut ni ne doit dépasser tels de modernes Procuste, ce bandit qui dans la mythologie proposait aux gens de les héberger avant que de leur couper les membres qui dépassaient de son lit.

Procuste dans l’histoire des plateformes c’est 140 signes sur Twitter puis 280 sur X, 2200 signes sur Instagram, 3000 pour LinkedIn, 150 signes pour une vidéo TikTok et 5000 signes pour décrire une vidéo Youtube. Malgré l’extension du domaine de la statusphère, Et tout le reste de ces espaces tant limités que limitrophes d’une atrophie des possibles.

(source : Dailyfy.co)

Comme le montre ce tableau, chaque espace est catégorisé par sa nature « titre de publication », « nom de la page », « description », « commentaire », « description vidéo », « bio », « DM message privé », etc. Et à chaque catégorie un autre espace, une autre limite : 58 caractères pour un titre de publication, 80 caractères pour une bio Tiktok, 30 pour un nom de profil Instagram, 1250 pour un commentaire LinkedIn, et ainsi de suite. Tout est métrique. A coups de trique. Quelques grands espaces demeurent : 110 000 caractères l’article LinkedIn mais 3000 caractères la « publication » de la même plateforme ; 63206 caractères le le post Facebook mais 8000 seulement pour le commentaire. Le numérique des plateformes est tout sauf l’espace infini qui était celui de la publication web des années 2000, c’est un espace totalement sous contrainte et en permanence sous astreinte.

Aujourd’hui l’essentiel de nos espaces discursifs en ligne qui ont recours au langage indépendamment de l’image ou de la vidéo sont des espaces Procustéens (images et vidéos par ailleurs elles-mêmes soumises à des limitations de durée). Des espaces où l’énonciation, où l’accès à la capacité de dire, ne vaut que par l’acceptation de cet horizon court de la diction. Certes il est possible, et certains y parviennent, de sublimer ces limites en les esthétisant ou en les poétisant. Les formes courtes ont par ailleurs une existence culturelle et littéraire longue. Mais nous parlons ici des conditions de production de la langue et donc du réel à l’échelle d’une humanité toute entière, en tout cas de sa part connectée (près de 5,5 milliards d’êtres humains). La « merdification » (enshittification) des plateformes, de presque toutes  les énonciations qui s’y déploient aujourd’hui, est aussi liée à cela. Ces espaces ne nous offrent pas tant la possibilité de dire qu’ils ne nous imposent la nécessité de rapidement nous taire. Il est d’ailleurs singulier de voir les plateformes parler de « liberté d’expression » dans des espaces aussi bornés et limités que ceux qu’elles nous offrent. Parlerait-on d’une « liberté de circulation » si celle-ci se limitait à un déplacement entre notre chambre et nos toilettes ?

D’un côté donc, il y eut, à l’échelle des plateformes numériques, la multiplication des espaces d’expression possibles. Mais en parallèle chaque nouvel espace énonciatif est apparu comme permettant de cadrer et de limiter « en nombre » toute expression singulière en ligne. Et désormais il y a la contamination de l’ensemble de ces espaces par des tiers énonciatifs logorrhéiques dopés à l’IA et aux artefacts génératifs. Voilà pour la parole « en émission ».

Quand à la parole « en réception », nous sommes aujourd’hui de plus en plus confrontés et exposés à ces dispositifs techno-logorrhéiques à proportion inverse de nos propres capacités à nourrir et à alimenter ces espaces expressifs et énonciatifs. Pour le dire plus trivialement : la part des bots et des contenus artificiellement générés explose et contamine l’ensemble des espaces numériques de l’ensemble des plateformes (c’est l’un des thèmes de mon livre « Les IA à l’assaut du cyberespace : vers un web synthétique » paru l’été dernier chez C&F Editions).

A l’échelle du numérique, si l’humanité se figeait à ce moment précis, et si nous l’observions à l’échelle macroscopique, nous serions alors forcés de constater que l’initiative de la parole, de la conversation, de la langue comme interaction et comme compréhension, que cette initiative revient davantage aux machines qu’à nous-mêmes. Et que notre réel s’y consomme d’abord, puis s’y épuise et s’y consume. Je le redis, chaque nouvelle simulation enlève au réel. Quand Trump annonce vouloir transformer Gaza en Riviera, des milliers d’images et de vidéos « génératives » viennent instantanément illustrer, montrer, filmer, documenter cet impossible, ce délire macabre autant que cynique. Ce qui était et aurait du rester impossible, y compris à formuler, devient sous le poids des générations illustratives un embranchement possible du futur qui nous est proposé comme instantanément disponible pour nos sens et nos imaginaires. C’est en cela qu’il enlève au réel, au seul réel possible qui est celui d’un arrêt des massacres en cours. Il ne s’agit pas uniquement d’élargir la fenêtre d’Overton mais bien de bâtir autour d’elle l’ensemble la totalité de la barre d’immeuble où elle ne sera plus qu’une fenêtre parmi d’autres, de faire exister la réalité qu’elle inaugure pour laisser au réel encore moins de chance de pouvoir la limiter ou la contraindre rapidement. C’est aussi cela que Trump a sinon compris, du moins intuitivement senti, et qui le fait tant à l’aise dans ces médias sociaux où prime l’expression courte et bouffonne : la parole politique, sa parole politique, se suffit au lit de Procuste tant que l’image y tient les dimensions voulues. Gaza en Riviera. Voilà ce qui seul doit tenir. Le reste s’abîme, s’épuise et s’effondre derrière ces dispositifs techno[-logo]rrhéiques pensés pour l’emballement viral et dont les IA et autres artefacts génératifs ne sont que la dernière – mais non ultime – instance contemporaine.

Les 3 chambres.

Depuis longtemps dans l’analyse des médias on utilise le concept de « chambre d’écho » pour désigner de manière métaphorique, le fait que « l’information, les idées, ou les croyances sont amplifiées ou renforcées par la communication et la répétition dans un système défini. »

Avec l’arrivée des médias sociaux, se fait jour chez Eli Pariser le concept de « bulle(s) de filtre« , cette idée que les algorithmes des moteurs de recherche comme des grands médias sociaux tendent à nous enfermer dans nos propres croyances et convictions.

Aujourd’hui, avec l’essor et la place que prennent les différents artefacts génératifs à l’identique de ChatGPT, des chercheurs (Jacob, Kerrigan, Bastos 2025) parlent d’un « Chat-Chamber effect ». Rien à voir avec les matous qui ont conquis l’internet, le « chat » doit être lu et compris comme le « Tchat », et caractérise l’un des biais de nos échanges conversationnels avec des artefacts génératifs, lequel biais désigne les informations incorrectes mais allant dans le sens du questionnement de l’utilisateur que les grands modèles de langage peuvent fournir ; des résultats et informations qui restent non contrôlées et non vérifiées par les mêmes utilisateurs mais auxquels ces mêmes utilisateurs font pourtant confiance. Le titre complet de leur article est ainsi : « L’effet ‘Chat-Chamber’ : faire confiance aux hallucinations de l’IA« .

Derrière ces 3 déclinaisons d’une même tendance à trois époques contemporaines successives, se jouent deux choses. D’abord des dynamiques médiatiques qui s’interpénètrent pour forger, favoriser et figer une certaine logique ou esthétique de réception des discours qui s’y tiennent. Ensuite une affordance informationnelle première qui est l’inverse du doute raisonnable : ces systèmes médiatiques ont pour affordance première leur capacité à nous faire croire qu’ils disent le vrai : les logiques virales de partage (« si c’est tant partagé c’est bien que ça doit être vrai« ), le dispositif de proximité (illusoire) avec l’énonciateur (« si c’est lui qui le dit alors cela doit être vrai« ), le biais de disponibilité (des informations délibérément erronées ou mensongères sont plus immédiatement accessibles et rendus visibles que d’autres plus sourcées) ne sont que quelques-uns des leviers de cette affordance de vérité supposée.

Depuis plus de 20 ans j’explique et documente le fait que les algorithmes sont des éditorialistes comme les autres, mais qui avancent essentiellement masqués et que les IA et autres artefacts génératifs installent dans une sorte de théâtre total, « gestes, sons, paroles, feus, cris« , dont plus personne n’est en capacité de comprendre les codes.

A.M.I. : Assemblage Machinique Informationnel.

L’arrivée des artefacts génératifs ajoute une dimension nouvelle et passablement problématique au tableau contemporain de la fabrication de nos croyances et adhésions. Jusqu’ici, moteurs de recherche et réseaux sociaux jouaient sur le levier déjà immensément puissant de leurs arbitraires d’indexation et de publication (le fait de choisir ce qu’ils allaient indexer et/ou publier) ainsi que sur celui, tout aussi puissant, de la hiérarchisation et de la circulation (viralisation) de ce qui pouvait être vu et donc en creux de ce qu’ils estimaient devoir l’être moins ou pas du tout.

Choisir quoi mettre à la « Une » et définir l’agenda médiatique selon le vieux précepte de « l’agenda setting » qui dit que les médias ne nous disent pas ce qu’il faut penser mais ce à quoi il faut penser. Ce principe premier de l’éditorialisation se double, avec les artefacts génératifs conversationnels à vocation de recherche, d’une capacité à produire des sortes d’assemblages machiniques informationnels, c’est à dire des contenus uniquement déterminés par ce que nous interprétons comme un « devenir machine** » en capacité de « phraser » les immenses bases de données textuelles sur lesquelles il repose. Des machines à communiquer mais en aucun cas, comme le souligne aussi Arthur Perret, en aucun cas des machines à informer.

** [ce « devenir-machine » est à lire dans le sens du « devenir-animal » chez Deleuze et Guattari : « le “devenir-animal” ne consiste pas à “imiter l’animal, mais d’entrer dans des rapports de composition, d’affect et d’intensité sensible” ».]

Ce concept d’assemblage machinique informationnel me semble intéressant à penser en miroir ou en leurre des agencements collectifs d’énonciation théorisés par Félix Guattari. D’abord parce qu’un assemblage n’est pas un agencement. Il n’en a justement pas l’agentivité. Il n’est mû par aucune intentionnalité, par aucun désir combinatoire, calculatoire, informationnel, communicationnel ou social. Ensuite parce que la dimension machinique est antithétique de la dimension collective, elle en est la matière noire : ChatGPT (et les autres artefacts génératifs) n’est rien sans la base de connaissance sur laquelle il repose et les immenses réservoirs de textes, d’images et de contenus divers qui ont été, pour le coup, assemblés, et sont l’oeuvre de singularités fondues dans un collectif qui n’a jamais été mobilisé ou sollicité en tant que tel. Enfin, l’information est ici un degré zéro de l’énonciation. L’énonciation c’est précisément ce qui va donner un corps social à ce qui étymologiquement, a donc déjà été « mis en forme » (in-formare) et se trouve prêt à être transmis, à trouver résonance. ChatGPT est l’ombre, le simulacre, le leurre d’une énonciation. Et cette duperie est aussi sa plus grande victoire.

La question du langage s’est déplacée. Un grand déplacement. Quelque chose qui n’est plus aligné entre le territoire du monde et la langue qui en est la carte.

Apostille(s). 

Pendant que nos réels sont saturés et épuisés d’IA, en Israël, un spot de publicité pour une entreprise d’armement utilise pour la première fois des images réelles d’une attaque de drone à Gaza. C’est à la fois la plus cynique et la plus perverse des formes de publicitarisation. Pendant ce temps Youtube dévoile un système baptisé « Peak Points » qui utilise son IA « Gemini » pour insérer des publicités aux moments où l’attention des spectateurs est supposément la plus forte grâce à l’exploitation de données « émotionnelles » (ou qui sont en tout cas supposées rendre compte de notre état émotionnel).

Couplez maintenant ces deux informations et déterminez sans algorithme votre propre état émotionnel …

De manière plus anecdotique, on constate un jaunissement par contamination de nombreuses images générées artificiellement. Les animaux étaient malades de la peste, les IA le sont déjà de consanguinité. Pendant ce temps encore, Youtube fait des retouches de vidéos sans en avertir les créateurs, pour « rapprocher esthétiquement le rendu des vidéos de ce que l’IA générative peut proposer. » Standardiser la consanguinité des contenus générés.

Pendant ce temps, des images pédocriminelles générées par IA sèment autant le trouble que le doute et entravent les démarches déjà si complexes du travail de police et de justice. La volumétrie des faux-positifs est ici la première alliée des pédocriminels. Et une nouvelle fois la charge de la preuve s’inverse comme si souvent dans l’histoire du numérique : il ne s’agit plus de pouvoir détecter la marginalité de faux-positifs sur un volume de contenus donnés, il s’agit de la maximiser pour que les résultats « sincères » soient marginalisés.

En 1985, dans un entretien où elle était interrogée sur sa vision de l’an 2000, Marguerite Duras prophétisait avec une rare acuité « Il n’y aura plus que des réponses. Tous les textes seront des réponses. (…) »  Nous sommes au moment où tout comme après l’avènement des moteurs de recherche au début des années 2000, il n’y a plus, aujourd’hui, que des réponses. De fausses réponses pourrait-on donc ajouter.

Marguerite Duras poursuivait : « Un jour un homme il lira. Et puis tout recommencera. » L’inflation des artefacts génératifs dans l’ensemble de nos écosystèmes informationnels et des sphères de nos savoirs sociaux (école, université, cercles politiques, etc.) pose autant la question de savoir ce qu’il restera encore à lire, que celle de savoir si nous serons en peine d’encore pouvoir écrire un destin commun. Christian Salmon le rappelait aussi dans sa récente analyse sur ce nouvel ordre narratif en train de s’imposer : « La prolifération de ces narrations non humaines interroge le devenir de l’humanité tout entière. »

La question du langage s’est déplacée, et notre technorrhée ressemble à s’y méprendre aux vers de Macbeth : un conte raconté par un idiot (ils sont en fait plusieurs), plein de bruit et de fureur, ne signifiant plus rien. (« It is a tale / Told by an idiot, full of sound and fury / Signifying nothing« ).

  • ✇The Kyiv Independent
  • The Kyiv Independent launches 'How to help Ukraine' newsletter
    The Kyiv Independent has launched a new membership benefit in direct response to the most frequently asked question from its members — "What more can I do to help Ukraine?"The "How to help Ukraine" newsletter will help members exercise their agency, get more involved in efforts to support the country, with donation suggestions and even volunteering opportunities, all while raising awareness of Ukrainian civil society.“We always knew that our community was passionate about supporting us, but that
     

The Kyiv Independent launches 'How to help Ukraine' newsletter

30 juin 2025 à 08:58
The Kyiv Independent launches 'How to help Ukraine' newsletter

The Kyiv Independent has launched a new membership benefit in direct response to the most frequently asked question from its members — "What more can I do to help Ukraine?"

The "How to help Ukraine" newsletter will help members exercise their agency, get more involved in efforts to support the country, with donation suggestions and even volunteering opportunities, all while raising awareness of Ukrainian civil society.

“We always knew that our community was passionate about supporting us, but that wasn’t the only thing driving them," Daryna Shevchenko, CEO of the Kyiv Independent said.

"Many of our members have repeatedly expressed a sincere desire to help Ukraine in ways that go beyond donating to the Kyiv Independent. Now, we’ve finally found the resources to act on that."

The first edition sent on June 12 raised $16,500 in just four days for the Leleka Foundation, a U.S.-Ukrainian charity that procures medical supplies for those on the front lines.

The Kyiv Independent conducted over a dozen in-depth interviews with supporters to find out exactly what they wanted, and the "How to help Ukraine" newsletter is the result — and it's already allowed members to give help where it's most needed.

The Kyiv Independent launches 'How to help Ukraine' newsletter
Screengrab from the Leleka Foundation website. (Leleka Foundation)

The first edition sent on June 12 raised $16,500 in just four days for the Leleka Foundation, a U.S.-Ukrainian charity that procures medical supplies for those on the front lines.

This money helped purchase medical backpacks for front-line medics containing tourniquets, bandages, hemostatic dressings, chest seals, soft splints, trauma shears, and burn treatment — each backpack has enough equipment to save 10 lives.

Ongoing support for organizations like Leleka Foundation is vital for Ukraine — it was founded in December 2014, when several Ukrainians in the U.S. joined forces with volunteer initiatives in Ukraine to procure and supply medical and first-aid equipment for paramedics, ambulances, and hospitals.

Leleka is currently a team of 20 staff members and volunteers across two continents and donations like those raised as a result of the "How to help Ukraine" newsletter will allow them to continue their work during uncertain times.

The "How to help Ukraine" newsletter will highlight one such organization each month, giving members the opportunity to donate money, time, or support to a group they can be assured shares the same values and mission as the Kyiv Independent.

"The desire to help Ukraine — more deeply and in more ways — is a passion we at the Kyiv Independent wholeheartedly share with our community," Shevchenko said.

For more information, please contact: news@kyivindependent.com

5 ways you can support Ukraine — even if your government doesn’t want to
Editor’s note: This is an edited version of a guide for our members published on Feb. 24, 2025. If you would like to join a community of people united by wanting to help Ukraine, consider becoming a member today. U.S. President Donald Trump’s public spat with President Volodymyr
The Kyiv Independent launches 'How to help Ukraine' newsletterThe Kyiv IndependentBrooke Manning
The Kyiv Independent launches 'How to help Ukraine' newsletter
  • ✇The Kyiv Independent
  • Ukraine to seek EU sanctions against Bangladesh over Russia-stolen grain import
    Ukraine is preparing to request European Union sanctions against Bangladeshi entities over importing wheat stolen from Russian-occupied Ukrainian territories, Ukraine's ambassador to India told Reuters.In comments for Reuters published on June 27, Oleksandr Polishchuk said that Bangladesh had failed to respond to Ukrainian diplomatic appeals to stop the trade, prompting Kyiv to escalate the matter to the EU level."It's a crime," Polishchuk told Reuters. "We will share our investigation with our
     

Ukraine to seek EU sanctions against Bangladesh over Russia-stolen grain import

27 juin 2025 à 08:08
Ukraine to seek EU sanctions against Bangladesh over Russia-stolen grain import

Ukraine is preparing to request European Union sanctions against Bangladeshi entities over importing wheat stolen from Russian-occupied Ukrainian territories, Ukraine's ambassador to India told Reuters.

In comments for Reuters published on June 27, Oleksandr Polishchuk said that Bangladesh had failed to respond to Ukrainian diplomatic appeals to stop the trade, prompting Kyiv to escalate the matter to the EU level.

"It's a crime," Polishchuk told Reuters. "We will share our investigation with our European Union colleagues, and we will kindly ask them to take the appropriate measures."

Russian forces have seized millions of tons of Ukrainian grain from occupied areas, with at least 180,000 tons looted through the port of Mariupol alone, Ukrainian Prime Minister Denys Shmyhal said in October 2024.

The grain theft from occupied territories began in 2014 when Russia first invaded Ukraine and continued after the full-scale invasion launched in 2022.

The Ukrainian Embassy in New Delhi sent multiple letters to Bangladesh's Foreign Ministry this year, requesting rejection of more than 150,000 tonnes of grain allegedly stolen and shipped from the Russian port of Kavkaz, according to documents reviewed by Reuters, shared by people familiar with the matter.

According to Polishchuk, Ukrainian intelligence showed that entities in Russia mask grain obtained from occupied Ukrainian territories by mixing it with Russian wheat before shipping it to other countries.

Bangladesh's Food Ministry, in comments to Reuters, denied the imports, saying that they banned grain deliveries originating from occupied Ukrainian territory.

Ukraine's agricultural sector has suffered approximately $80 billion in losses due to Russia's full-scale invasion, Agriculture Minister Vitalii Koval said in February.

The losses include direct damage, disrupted logistics, higher fertilizer and fuel costs, land reclamation expenses, demining costs, and impacts from occupied territories.

Amid relentless Russian strikes, Ukraine’s businesses rebuild alone
It took firefighters two days to extinguish the flames at Oleksiy Tarnopolskiy’s warehouse in Kyiv after a Russian attack on June 10. Nothing inside could be saved from his tea and coffee business. Tarnopolskiy arrived on site to an “apocalyptic” scene at 5 a.m., one hour after a Russian
Ukraine to seek EU sanctions against Bangladesh over Russia-stolen grain importThe Kyiv IndependentDominic Culverwell
Ukraine to seek EU sanctions against Bangladesh over Russia-stolen grain import
  • ✇The Kyiv Independent
  • The Kyiv Independent reaches 20,000 paying members — thanks to our global community
    The Kyiv Independent has officially reached 20,000 paying members, becoming the first Ukrainian media outlet in history to cross this milestone.This is  not just a number — it’s a reminder that people around the world believe in the importance of independent journalism from Ukraine, and are willing to stand behind it.When Russia launched its full-scale invasion in February 2022, the Kyiv Independent knew how important it would be to keep its reporting free and accessible — to show the world what
     

The Kyiv Independent reaches 20,000 paying members — thanks to our global community

26 juin 2025 à 11:33
The Kyiv Independent reaches 20,000 paying members — thanks to our global community

The Kyiv Independent has officially reached 20,000 paying members, becoming the first Ukrainian media outlet in history to cross this milestone.

This is  not just a number — it’s a reminder that people around the world believe in the importance of independent journalism from Ukraine, and are willing to stand behind it.

When Russia launched its full-scale invasion in February 2022, the Kyiv Independent knew how important it would be to keep its reporting free and accessible — to show the world what was really happening on the ground.

More than three years later, that mission hasn’t changed. That´s why, in May, the Kyiv Independent launched a campaign to ensure that no matter what happens, it would be able to continue showing the threat Russia poses to the world, and the consequences of when it is allowed to do so unchecked.

In less than one month, 2,400 people became new members of the Kyiv Independent community.

  • Support came from more than 60 countries, with the highest number of new members joining from North America and Europe.
  • To help members feel more connected, an updated interactive community map was unveiled — since the update, over 700 people have added their pins to show where they're supporting from.
  • A live campaign counter allowed every person who joined to see their impact in real time — and know they were contributing to something bigger.
  • Visitors to the Kyiv Independent website saw country-specific messages during the campaign — small touches of language or culture that helped people feel seen, no matter where they were coming from.

Since its founding in late 2021, more than 50,000 people from over 130 countries have contributed financially to support the Kyiv Independent — either through one-time donations or recurring membership.  In 2024, reader contributions accounted for 70% of revenue.

The Kyiv Independent reaches 20,000 paying members — thanks to our global community
A map of The Kyiv Independent supporters from around the world. (The Kyiv Independent)

“The journey to twenty thousand took us 3.5 years and thousands of hours of work by the most talented people I know. It’s a testament to our journalism and our mission to provide high-quality reporting from Ukraine to the world.” said Zakhar Protsiuk, COO of The Kyiv Independent. “These twenty thousand people are our secret power that allows us to do our work and grow.”

All of this is a powerful reminder of what a committed community can do, and how reader-funded journalism can thrive, even in the hardest conditions.

The team at the Kyiv Independent is incredibly grateful to everyone who joined, gifted a membership, or shared the campaign. This achievement belongs to all of them — and it shows what’s possible when journalism is powered by community.

For more information, please contact: news@kyivindependent.com

  • ✇The Kyiv Independent
  • EU approves new tariffs on Russian, Belarusian agricultural goods
    The Council of the EU on June 12 approved fresh tariffs on fertilizers and remaining agricultural goods from Russia and Belarus, aiming to reduce Russian export revenues.The measures target those goods that have not yet been subject to additional customs duties and will enter into force on July 1. The tariffs on fertilizers will increase gradually over the next three years.The step comes as the EU readies additional sanctions against Russia as it continues to wage its all-out war against Ukraine
     

EU approves new tariffs on Russian, Belarusian agricultural goods

12 juin 2025 à 08:42
EU approves new tariffs on Russian, Belarusian agricultural goods

The Council of the EU on June 12 approved fresh tariffs on fertilizers and remaining agricultural goods from Russia and Belarus, aiming to reduce Russian export revenues.

The measures target those goods that have not yet been subject to additional customs duties and will enter into force on July 1. The tariffs on fertilizers will increase gradually over the next three years.

The step comes as the EU readies additional sanctions against Russia as it continues to wage its all-out war against Ukraine.

"Polish Presidency motto is 'Security, Europe!' and these measures increase our economic security by reducing dependencies from Russia," said Michal Baranowski, the trade undersecretary at the Polish Economy Ministry.

"We are further reducing Russia’s export revenues and therefore its ability to finance its brutal war. This is united Europe at its best," he said in a statement.

The new tariffs will apply to goods that made up around 15% of all agricultural imports from Russia in 2023. Fertilizer tariffs will focus on certain nitrogen-based products, the Council said in a statement.

Russian fertilizers accounted for more than a quarter of all of the EU's imports in this sector in 2023, worth almost $1.5 billion.

Apart from stifling Russia's trade revenue, the step also aims to reduce the EU's dependence on Russian and Belarusian goods, protect European farmers, and diversify the supply.

The EU adopted higher tariffs on cereals, oilseeds, and some other products from Russia and Belarus in May 2024. Earlier this year, the European Commission proposed imposing similar measures on all remaining agricultural products from the two countries.

Putin’s suspected daughter found working in anti-war galleries in Paris
Nastya Rodionova, a Russian writer and artist who has been based in Paris since 2022, had only met gallery manager Luiza Rozova in passing at events before she learned who the 22-year-old’s parents were. Described by a number of people as a “very nice and well-mannered girl,” Rozova is
EU approves new tariffs on Russian, Belarusian agricultural goodsThe Kyiv IndependentKate Tsurkan
EU approves new tariffs on Russian, Belarusian agricultural goods
  • ✇Made Not Found (by danah boyd)
  • Welcome to my newsletter!
    Welcome to Made Not Found, the newsletter that I (danah boyd) write to share random thoughts, ideas, and updates. Much of the content here is also posted on Apophenia, my blog.If you'd like to join me here, please subscribe and you'll get emails when new content is published!
     
❌