EU classifies ChatGPT as a very large search engine

ChatGPT now sits in one of the EU’s toughest regulatory categories for online services. Its classification as a Very Large Online Search Engine under the Digital Services Act puts OpenAI under tighter oversight for risks involving minors, users’ mental health and the circulation of illegal content.

The EU will treat ChatGPT as a major online service covered by its rules for large search platforms. OpenAI must mitigate specific risks linked to how people use the chatbot. The supplied facts do not describe the designation as a ban or identify any immediate restriction on access to ChatGPT.

The classification moves scrutiny beyond individual answers to the service’s broader effects. The stated risk areas include harms involving young users, mental health and illegal material. Analysis: placing ChatGPT in a category built for major online services means regulators can assess how OpenAI addresses those risks across the service.

OpenAI is directly affected. EU users are affected through the safeguards applied to ChatGPT in the bloc. Minors receive particular attention in the identified risk areas, while the focus on mental health pushes the issue beyond conventional content moderation.

Illegal content creates a separate challenge. ChatGPT generates responses through conversation instead of presenting a conventional list of search results. Applying search-engine rules to that interface is like fitting an existing traffic code to a new kind of vehicle: the safety goal stays the same, but regulators and OpenAI must translate it into controls suited to the technology.

Based on the supplied facts, compliance centres on three areas:

• Mitigating risks connected to ChatGPT’s impact on minors.
• Addressing risks to users’ mental health.
• Reducing the spread of illegal content through the service.

The source does not specify which technical measures OpenAI must introduce. It also does not identify reporting requirements, enforcement steps or potential penalties. The designation places OpenAI under tougher obligations, but the operational burden will depend on how those obligations are applied.

The report was published on August 31, 2026, and said OpenAI would face the tougher obligations soon. The supplied facts give no exact compliance deadline or separate effective date. The classification is confirmed; the implementation schedule is not.

Analysis: the practical question is which controls OpenAI will introduce for minors, mental-health risks and illegal content, and when those requirements will take effect.