The principle of Grok 2 is simple: type a few words and the AI instantly creates photos that are more real than life. Emmanuel Macron shaking hands with Mussolini in front of Notre-Dame? Child’s play. Donald Trump and Jean-Luc Mélenchon in spacesuits? Done in seconds.
Unlike its competitors DALL-E or Midjourney which favor artistic rendering, Grok 2 focuses everything on photorealism. Even more worrying, Musk’s AI imposes virtually no ethical restrictions. While ChatGPT categorically refuses to generate images of public figures, Grok 2 does so without flinching.
This absence of safeguards is not an oversight but a deliberate strategy which is part of Elon Musk’s libertarian vision. The billionaire, who has already transformed Twitter into a temple of disinformation, is doing it again with an AI capable of creating any compromising situation involving celebrities. Only nudity remains prohibited.
“Its free nature lowers the threshold of technicality required to generate false realistic images”warns Tristan Mendès France, specialist in digital cultures, to the Figaro. With 25 images generated every two hours for free, X’s 600 million users can now produce a continuous stream of misleading content.
https://x.com/ViolonViande/status/1868377074622439655
Threat to democracy
The timing couldn’t be more concerning. While 2024 was a year of major elections in many countries, particularly in the United States, the arrival of Grok 2 raises fears of an explosion of disinformation campaigns for those to come. Western intelligence services have already identified organized networks, such as “Portal Kombat”, which distribute pro-Russian content to destabilize public opinion. The Romanian elections, in which the far-right emerged victorious, were, for example, largely influenced by TikTok, the authorities explain.
Elon Musk’s tool now offers them a formidable weapon. NewsGuarda company specializing in the analysis of disinformation, has already identified 767 fake news sites generated by AI, including 15 in French. With the integration of Grok 2 images, these platforms will gain credibility and virality.
For the moment, a few imperfections still betray the artificial origin of the images: strange looks, suspicious blurring effects, illegible texts… But these defects could soon disappear. Elon Musk has already announced a version 3 of Grok with computing power “10 to 20 times greater”. AI could even soon render “extremely compelling legal verdicts”according to its creator.
Faced with this growing threat, experts and regulators are sounding the alarm. Without strict supervision, Grok 2 could well become the lethal weapon of disinformation merchants, further undermining confidence in online information, which is already seriously damaged.
- Grok 2, Elon Musk’s AI, generates ultra-realistic photos for free with almost no ethical restrictions
- Unlike its competitors, it can create images of public figures in any situation
- Its ease of use and its free nature make it the ideal tool for massive disinformation, particularly worrying in an already tense geopolitical context.
???? To not miss any news from Presse-citron, follow us on Google News and WhatsApp.
