Russia's invented Polish soldiers, a fake Hezbollah threat for Bastille Day, and AI mourners at Khamenei's funeral

Russia fakes Polish soldiers to split Poland and Ukraine

Russia is using AI to build soldiers who do not exist, and this month it aimed them at Poland and Ukraine. The two neighbours are already at odds. After a Ukrainian military unit was named after a wartime nationalist group that Poland associates with the 1943 Volhynia massacres, Poland's president stripped President Zelensky of Poland's highest state honour. Russia moved straight into that gap.

Poland's state research group, NASK, found AI-made videos of fake “Polish soldiers.” In them, the invented figures attack Ukraine and push a false claim that Ukrainian refugees have driven up crime in Poland. They also reach back to the Volhynia massacres themselves, when wartime Ukrainian nationalist forces killed tens of thousands of Poles. It is a real and still-sensitive chapter of the war, and Russia keeps reopening it to drive Kyiv and Warsaw apart.

The giveaways pointed back to Russia: NASK found the scripts had been translated word-for-word from Russian into Polish, with stray Cyrillic letters left in, and the uniforms looked closer to Russian or Belarusian kit than to Poland's. The most-viewed clip reached nearly 300,000 views before several accounts were taken down, and the Polish fact-checking outlet Konkret24 found one YouTube account alone had posted more than 50 of the videos.

The same playbook targets Ukrainians directly. One fake Ukrainian soldier claimed his comrades die in the war so politicians can buy villas on the French coast, drawing about 915,000 views on Facebook. Around the same time, Russia fed the TV channel RT fake “secret documents” about that 1943 massacre, and Poland's government answered them point by point.

Researchers say the goal is not really to make anyone believe a single clip; it is to spread confusion and wear down trust, both in Ukraine's army and between Warsaw and Kyiv.

A fake video threatens France on Bastille Day

Just before July 14, a video spread on X. It showed three masked men in camouflage standing over a dummy dressed as a French soldier, its face painted with the French flag. They threatened to “shed blood” on Bastille Day if France kept arming Israel, and one of them beheaded the dummy with a knife. It was styled as a message from Hezbollah, the armed group in Lebanon, and it reached almost a million views.

The timing was the point: France was about to hold its July 14 military parade, which President Macron had turned into a show of strength while war continues in Europe. A staged terror threat aimed at that exact day was built to unsettle.

The video was fake. Euronews checked it: a real Hezbollah message would carry the group's logo, and this one had none. The accent was Levantine but not Lebanese, and the speaker made basic grammar mistakes. It did not cause the panic it was designed for. An expert asked by AFP called it not very credible, and in the comments many users doubted it, with some guessing it was staged.

Researchers link the clip to Storm-1516, a Russia-linked operation that pumps out fabricated videos and pushes them through friendly accounts. It is the same network blamed for earlier fakes smearing Macron.

Viral AI fakes flood social media as Iran mourns Khamenei

When Iran buried its leader, Ayatollah Khamenei, large crowds really did turn out in Tehran. The trouble was the footage that spread fastest online, much of it fake.

France 24's fact-checkers spotted the signs. Some images had made-up text that looked like Persian but meant nothing. In one, a building changed colour from one frame to the next. One AI photo even carried a hidden watermark that flagged it as AI-made. One fake video of the crowds drew hundreds of thousands of views, with the usual AI giveaways on show.

The fakes served two ends: some were just chasing clicks off a big story, which France 24 says is common; others were deliberate, with regime supporters using AI to make the crowds look bigger while opponents pushed fakes the other way.

China and Russia push anger over data centers

Many Americans do not want new data centers built near their homes, and the anger crosses party lines. It has pulled in figures as far apart as Bernie Sanders and Steve Bannon. State actors in China, Russia and Iran have moved to make that anger worse.

Researchers counted about 700 mentions of data centers in Chinese, Russian and Iranian state media in the first six months of the year, roughly four a day. A Chinese state newspaper ran a satellite photo of a data center in Gainesville, Virginia, warning in English that AI threatened Americans' health and their wallets. A covert Russian operation questioned a US-built data center in Armenia, a country under heavy Kremlin pressure. Chinese accounts also used ChatGPT to make cartoons blaming data centers for soaring electricity bills, and OpenAI reported similar AI-made efforts, including a fake Maryland comic strip that drew almost no real engagement.