EU may treat ChatGPT as platform

The European Commission said on April 10 that it is analyzing whether OpenAI’s ChatGPT should be treated as a very large online platform or search engine under the Digital Services Act. (msn.com) That question matters because the Digital Services Act sets its toughest tier at more than 45 million average monthly users in the European Union. OpenAI says ChatGPT search had about 120.4 million average monthly active recipients in the bloc for the six months ending September 30, 2025. (digital-strategy.ec.europa.eu, help.openai.com) If Brussels formally designates ChatGPT under that regime, OpenAI could face annual independent audits, public risk assessments, and closer Commission supervision on top of rules it already faces as an artificial intelligence provider. (digital-strategy.ec.europa.eu, digital-strategy.ec.europa.eu, digital-strategy.ec.europa.eu) The overlap would land as the European Union is still phasing in its Artificial Intelligence Act. The law entered into force on August 1, 2024, banned certain uses from February 2, 2025, applied general-purpose model rules from August 2, 2025, and brings most high-risk system rules in from August 2, 2026. (digital-strategy.ec.europa.eu, digital-strategy.ec.europa.eu, eur-lex.europa.eu) The two laws do different jobs. The Digital Services Act regulates online intermediaries and search services at scale, while the Artificial Intelligence Act regulates how artificial intelligence systems and general-purpose models are built, documented, tested, and placed on the market. (digital-strategy.ec.europa.eu, digital-strategy.ec.europa.eu) For high-risk artificial intelligence, the Artificial Intelligence Act requires a conformity assessment before a system can carry a CE mark and enter the market. The law says outside “notified bodies” may be needed when harmonized standards are missing or only partly applied. (artificialintelligenceact.eu, ai-act-service-desk.ec.europa.eu) That is where some lawyers and compliance specialists see friction: one rulebook can ask whether a service creates systemic platform risks, while the other asks whether a model or system meets product-style safety and documentation rules. The Commission has already said most high-risk Artificial Intelligence Act obligations do not start until August 2, 2026, with some provisions extending to August 2, 2027. (digital-strategy.ec.europa.eu, eur-lex.europa.eu) OpenAI has already been publishing Digital Services Act user numbers and a government contact point for European authorities. The company’s help page says those figures were calculated for Digital Services Act compliance purposes and “should not be relied upon” for other uses. (help.openai.com) The Commission has not yet added ChatGPT to its public list of designated very large online platforms and search engines. Until that changes, the story is less about a final ruling than about whether Brussels decides ChatGPT is no longer just an artificial intelligence product, but also a platform it must police at internet scale. (digital-strategy.ec.europa.eu, msn.com)