EU labels ChatGPT a search engine. Disclosure rules follow.
DSA compliance will force OpenAI to describe how ChatGPT retrieves and ranks information, giving B2B brands their first regulated window into LLM citation logic.
Key takeaways
- The EU Commission has classified ChatGPT as a very large search engine under the DSA, with compliance required by end-2026.
- OpenAI must publish risk assessments, transparency reports, and an ad archive, the first structured disclosures about ChatGPT's retrieval mechanics.
- An advertising archive requirement means any future commercial content integrations in ChatGPT answers will be publicly visible.
- Whether the Commission can access OpenAI's training data remains legally disputed; a ruling in favour would reveal why certain brands appear in answers more reliably.
- B2B brands in financial services, industrials, and multilaterals should treat EU AI answer visibility as a compliance-adjacent discipline, not an informal content exercise.
The European Commission has classified ChatGPT as a very large online search engine under the Digital Services Act, The Decoder reports, triggering a compliance clock that expires at the end of 2026. OpenAI must deliver risk assessments, transparency reports, and an advertising archive. The threshold for the designation is 45 million monthly active users in the EU; ChatGPT clears it.
The classification matters less for what it demands of OpenAI and more for what it will reveal about how ChatGPT surfaces information. Transparency reports under the DSA are not press releases. They are structured disclosures about systemic risks, content moderation, and, crucially, the mechanics of how results are generated and ranked. For the first time, brands operating in the EU will have a legally mandated paper trail to scrutinise.
The compliance obligations are not symmetrical
Google and Bing have lived under DSA very-large-search-engine rules since 2023. Their transparency reports now run to hundreds of pages, covering algorithmic systems, recommender logic, and risk mitigation measures. ChatGPT's disclosures will need to match that standard. The difference is that Google's index is legible in principle: crawlable, linkable, auditable with third-party tools. ChatGPT's citation behaviour is not. It has no public index, no canonical ranking signal, and no clear mechanism by which a brand can confirm its content was retrieved and used.
DSA compliance will not fix that opacity entirely, but it creates a legal obligation to describe the system. OpenAI will have to characterise, at minimum, how ChatGPT retrieves and presents information in response to queries the Commission considers search-like. That characterisation, however partial, is more than the market currently has.
For senior marketers at financial institutions, industrial groups, and multilateral bodies, this is consequential. The EU accounts for a significant share of high-value B2B decision-making. An analyst at a European bank asking ChatGPT about credit risk frameworks, a procurement officer at a major industrial asking about supplier certification standards, a policy researcher at a UN agency asking about climate finance instruments: these are live use cases. The sources ChatGPT cites in those answers shape institutional knowledge. Until now, the selection logic has been entirely opaque.
What the ad archive changes
The requirement for an advertising archive is the sleeper provision. ChatGPT does not currently run display advertising, but OpenAI has not ruled out commercial content integrations. The DSA archive requirement covers any "advertisement presented on the interface." If OpenAI introduces sponsored placements, featured answers, or commercially influenced retrieval, the archive makes those visible. For brands that have no formal relationship with OpenAI but appear (or fail to appear) in answers, knowing whether commercial arrangements shape visibility is not a theoretical concern.