“GTA 6 ALERT – EXTREMELY SERIOUS SITUATION”. This message, published on X (formerly Twitter) by the Zap Actu GTA6 account, accompanies a short video sequence supposed to show a gameplay scene from the highly anticipated Grand Theft Auto VI. Very quickly, the post went viral, accumulating 8 million views in just 24 hours. Except it’s all false.
The video is not a leak. It’s not even a real game capture, but a simulation created by artificial intelligence. This is not an isolated case: other similar content published by this same account have also exceeded one million views in recent months. The objective is clear: generate traffic, gain subscribers, and attract members to their Discord server.
These deepfakes are increasingly convincing, to the point of fooling a large part of the public. Some users, ill-informed or overly enthusiastic, take these videos at face value. A worrying phenomenon, which does not only affect Rockstar and its cult franchise.
AI deepfakes: a threat that extends well beyond video games
The GTA 6 case is emblematic, but far from unique. For months, celebrities and public figures have denounced the abusive use of their image for commercial purposes or disinformation. Actor Keanu Reeves declared that “it’s not very funny” upon discovering deepfakes of him used to sell products without his authorization.
In 2023, Tom Hanks also warned his fans against an advertisement for dental care exploiting his image via AI, without any agreement. Morgan Freeman thanked his audience after being alerted to imitations of his voice, also generated by artificial intelligence. As for Jamie Lee Curtis, she publicly called out the boss of Meta, after seeing an advertisement broadcast with her image for “some bullshit that I didn’t authorize, agree to or endorse”.
Even physicist Brian Cox was not spared: he recently denounced fake videos in which he made absurd remarks about a comet. Each time, the process is the same: powerful technology, viral content, and no validation from the people involved.
GTA 6, OpenAI, and the blurred limits of “fan” creation
Since October 1, the Sora 2 tool launched by OpenAI allows you to generate 1080p videos of 20 seconds, with sound. Very quickly, sequences invaded social networks, featuring characters from protected universes like MarioPokémon, One Piece or Demon Slayer. In response, the Japanese government officially asked OpenAI to respect copyright.
The firm’s CEO, Sam Altman, downplayed the accusations, believing that these creations amount to “interactive fan fiction” – interactive fiction created by fans. A legal and artistic vagueness that suits viral content creators well.
On Discord, some people are still wondering. A user directly challenged the Zap Actu GTA6 account in these terms: “If it’s AI, that’s disrespectful to the fans, friend @ZapActu.” The account did not respond. He preferred to publish a new video, which already had more than 220,000 impressions just a few hours after it was posted online.