Fake rabbit on the trampoline. Italian brain rot trend. Fake coup in France. Latest Coca-Cola and McDonald's advertisements. Donald Trump propaganda clip. Tivo in shape [French fitness influencer] Make fun of racist posts featuring yourself. All these videos are generated by artificial intelligence and have made headlines in recent months. But AI-generated content has been quietly fueling the daily feeds of most social media users, often without them even realizing it.
The signs are everywhere. In the spring, we analyzed the 10 most popular YouTube channels in the US and found that four of them systematically posted AI-generated videos, primarily music content. In November, a TikTok spokesperson said there were 1.3 billion such videos verified on the platform.
At the root of this wave is the democratization of generative AI video creation tools. These tools – Veo (launched by Alphabet), Sora (OpenAI), CapCut (Bytedance), MovieGen (Meta), Grok (xAI) – are now directly integrated into consumer AI platforms (Gemini, ChatGPT, Meta AI, etc.). With simple written prompts, anyone can create a video in minutes. And in just a few months, the realism of these videos has improved dramatically.
Most of the companies behind these technologies also control the most popular social networks (YouTube, TikTok, Facebook, Instagram, X). According to Alexios Mantzaris, a researcher specializing in disinformation at the site Indicator, these companies need to prove that AI has a purpose, which is why they put these tools into the hands of users. Mantsalis believes the move is aimed at legitimizing AI-driven valuations amid growing concerns about a potential tech bubble.
“Sell advertising”
67.69% of this article remains. The rest are exclusive to subscribers.
