Brussels just put ChatGPT on notice. The European Commission classified OpenAI’s chatbot, alongside Reddit and Roblox, as a very large online platform under the Digital Services Act, which means the three services now face obligations to remove illegal content, protect minors’ privacy and security, and potentially get hit with fines up to 6 percent of global revenue if they don’t comply. All three crossed the 45 million monthly active users threshold in the EU that triggers the enhanced scrutiny.
EU tech chief Henna Virkkunen didn’t mince words. “ChatGPT, Reddit, and Roblox will now be held to a higher standard of scrutiny and accountability in the European Union, in line with their large impact on our citizens and society,” she said. The clock is already ticking: they have until the end of December to meet the additional requirements.
This isn’t the first time the DSA has reached into generative AI. X’s Grok chatbot is already under investigation under the same law, and the Commission’s move signals that Brussels sees no meaningful distinction between traditional social platforms and AI services when it comes to content moderation and user safety. The 45 million user threshold, originally designed with social networks in mind, is now functioning as a tripwire for AI adoption across the bloc.
What’s striking here is the speed at which ChatGPT hit that number. It took Facebook years to reach 45 million EU users; ChatGPT did it in roughly two. That trajectory likely means other generative AI tools aren’t far behind, and the Commission’s decision creates a template for how it will treat them. For OpenAI, the real question is what “removing illegal content” means in practice for a model that generates text on demand rather than hosting user posts. That ambiguity is where the compliance headaches — and the lawyers — will live.
