EU puts ChatGPT under the Digital Services Act's strictest tier
Two regulatory fronts moved at once today: Brussels pulled the world's most-used chatbot into its toughest platform regime, and the US state-level AI fight split the labs down the middle. Neither is the headline-grabbing kind of AI news — no model launch, no benchmark — but both will shape what ships and where.
The European Commission has designated ChatGPT a Very Large Online Search Engine and Reddit and Roblox Very Large Online Platforms under the Digital Services Act, pulling all three into the bloc's strictest tier for digital services. The threshold is 45 million monthly EU users, and the roster now runs to 28 platforms and search engines. Companies get four months to comply; miss the deadline and the exposure is fines of up to 6% of annual global revenue.
The designation is the interesting part, not the fine. Brussels classified ChatGPT as a search engine rather than as an AI service — which is the only hook the DSA currently offers. OpenAI spokesperson Oscar Haines said ChatGPT search "operates as a search service under the DSA," and that the company is preparing for the additional compliance duties. Roblox's European general manager Joost Hagesteijn called it a milestone and "the first gaming platform to reach" it. Reddit said it had been preparing too.
That framing is where the real story sits. Regulating ChatGPT as a search engine covers the search surface and leaves the chat itself largely untouched — the conversational product that hundreds of millions of people actually use is not what the designation reaches. Our own take: this is Brussels fitting a new technology into a 2022 statute because it has nothing better to hand, and everyone involved knows it. The compliance burden is real, but the gap between what the DSA covers and what ChatGPT does will be the thing to watch when the first enforcement action lands.
Anthropic and OpenAI are now on opposite sides of the US state AI regulation fight, and the split is widening. Massachusetts is advancing what has been described as the nation's most stringent state-level AI safeguards — stricter than the laws already adopted in California, New York and Illinois. The provisions sit inside a larger economic development bill that passed the state Senate last month and still needs House and gubernatorial approval, with negotiators working through the details.
The mechanism is what sets it apart. Illinois requires a third-party audit once a year to check that labs follow their own safety guidelines. The Massachusetts bill would let outside organizations evaluate a model's potential dangers every 120 days against those organizations' own criteria, and publish what they find. The state could not halt AI development — but publicity is the enforcement lever, and it is a sharp one.
The strategic read matters more than the clause. Anthropic has been backing state-level bills while OpenAI pushes a single federal standard that would pre-empt state rules. Both companies are heading toward IPOs, which makes this a disclosure question rather than a pure policy one: a lab whose existing safety framework already satisfies the state requirements absorbs the cost of compliance far more easily than one that has to build the infrastructure from scratch. That asymmetry, not the safety argument, is what is driving the divergence.
Which lever do you think actually changes lab behavior — Brussels' fines, or Massachusetts' 120-day public audits? Tell us in the comments.
