Video shared with claims it shows a large ‘pro-Palestine’ demo in Japan is AI

10 minute read

Author

Date

18/8/2025

Share

A video shared on social media with claims it shows a large demonstration “for Palestine” in Japan has been created using artificial intelligence (AI).

The footage has been circulating on X (formerly Twitter) and Facebook with claims it is the second largest demonstration of its kind after Sydney, Australia. It appears to feature two clips, one showing thousands of protesters on the street of a city, and another showing crowds outside a stadium-style building.

In the second half of the clip, Palestinian flags are present, while in the first, signs and banners with what appears to be Japanese lettering can be seen.

The AI-generated video being shared on Facebook.

The first clip also includes audio of what appears to be a crowd chanting “from the river to the sea, Palestine will be free”, while the second includes an unintelligible crowd chant.

But experts have confirmed that the footage isn’t real and has been made using AI.

Professor Hany Farid, who specialises in digital forensics, misinformation and image analysis at the University of California, Berkeley, and is Chief Science Officer at GetReal Security, a cybersecurity company focused on preventing malicious threats from generative AI, told us there were “several tell-tale signs that this video is AI generated”.

These include the fact that each clip is under eight seconds long, which is “typical” of current video generators. The motion in the crowd also “has the tell-tale signs of being AI generated” and Mr Farid’s team extracted a SynthID watermark from both the audio and video, showing that this video was made with a Google product. Mr Farid’s team suspect this was most likely Google’s AI video generator, VEO 3.

Google has previously told us that content produced with its AI products contains an embedded ’SynthID’ digital watermark that is not visible to the human eye.

This watermark can be used to identify content which comes from Google’s AI tools, though Google told us the presence of a watermark can't tell us whether AI was used to completely generate a brand-new image or just modify an existing one.

When we reverse image searched a frame from the video using Google, it also told us that it was “Made with Google AI”.

Emmanuelle Saliba, Chief Investigative Officer at GetReal Security, added: “Contextually, there is no geolocation match. The stadium-like structure doesn't correspond to any of the known protest sites in Japan.”

She also noted that “some of the protest signs and building signage are visibly scrambled”, which is “a common GenAI artifact”.

There have been reports of protests about Palestine in Japan in recent months, but the viral video does not show these.

It is also difficult to accurately estimate crowd sizes at large unticketed events, such as protests. While pro-Palestinian protests have taken place in Sydney, Australia, in August, we could see no evidence that these were the largest of their kind worldwide.

We have previously fact checked other misleading content about protests around the world. Before sharing media that you see online, first consider whether it comes from a verifiable and trustworthy source. Our guides to spotting AI-generated content can help you do this.