VERDICT: FALSE — The viral photo of an elderly couple stripped to their underwear by rioters in Paris is AI-generated. No such attack was reported to French prosecutors, and the person who made the image has admitted creating it with ChatGPT.
The claim
One of the most shared images of France’s October unrest shows an older man in his underwear and an older woman in a slip, standing on a Paris street surrounded by masked, dark-clad figures. Posts carrying it — on X, Facebook and Instagram — claimed the couple, said to be in their eighties, had been stripped of everything by rioters near Place de la République in Paris, with some captions blaming “far-left and migrant rioters.”
The image travelled fast and far. One version stripped of its original AI label drew more than six million views on X alone, and it was picked up by police-union accounts, politicians and television commentators abroad, presented as proof of how bad the French unrest had become. As recently as 23 October, Facebook posts were still sharing it as a genuine photograph of a real attack.
France has indeed seen serious unrest: mass protests over living costs and school conditions since late September, culminating in a national day of action on 17 October that brought looting and clashes in several cities. That real violence is precisely what makes a fake image like this so effective — it slots neatly into something readers already believe is happening.
The facts
First, the image itself gives the game away. Fact-checkers who ran it through Google’s SynthID detector found the invisible watermark carried by AI-generated images, indicating it was made or edited with Google’s AI tools. The picture is also full of the small errors typical of generated imagery: a masked figure with two heads on one body, a woman’s hand with a missing finger, and figures in the background whose bodies merge into each other.
Second, and decisively, the image’s creator has owned up to making it. BBC Verify traced the picture to Infos-Israel.News, a small news site, whose director, Sabrina Hania, told the BBC her team created it using ChatGPT, without any reference photograph, and published it on 17 October. She said she regretted that it had been taken out of context and had misled so many people, blaming those who reshared it without its AI label. The original post, notably, was labelled “made with AI” on X and “AI content” on Instagram — the labels were cut off when others copied and reposted it.
Third, there is no official record of the attack the image supposedly documents. The Paris public prosecutor’s office told the French newspaper Libération’s CheckNews desk on 19 October that it had “received no reports to that effect to date.” French prosecutors repeated the point in the days that followed: no complaint, no report, no case matching the story.
How did the story start? On the evening of 17 October, a television journalist reported on air, citing a police source, that an elderly couple had been attacked near Place de la République. No image was shown in that report, and the alleged incident itself has never been confirmed. Hours later, the AI image appeared online, illustrating a claim that had no photographic evidence — and readers understandably assumed the picture was that evidence. This is the second AI image from the French unrest debunked on this site — the first was a viral video supposedly showing flags hung from the top of the Eiffel Tower.
Verdict
FALSE. The photo is not a photograph of a real event. It was generated with ChatGPT, published with an AI label that later reposters removed, and the attack it purports to show was never reported to prosecutors. The underlying unrest in France was real, but this particular outrage never happened.
Why it matters
A fabricated image like this does damage in both directions. It invents a crime against two vulnerable people who do not exist in this story, and it hardens a false picture of who the protesters are and what they did — in this case, feeding claims about “migrant rioters” that the evidence does not support. It also poisons the well for real victims: when the next genuine report of violence during a protest emerges, readers primed by fakes like this one will be quicker to doubt it. The lesson of this image is a simple one, and it cuts against the grain of how social media works: the more perfectly a picture confirms what you already believe, the more carefully it deserves to be checked.
