ChatGPT is no longer merely competing with search engines for users’ questions. In the European Union, it is now officially regulated as one.
On August 31, the European Commission designated ChatGPT as a Very Large Online Search Engine, or VLOSE, under the Digital Services Act. The decision places OpenAI’s service inside the EU’s highest regulatory tier for online search services, alongside obligations designed for products whose scale can create systemic effects across the European information environment.
The designation formalizes a shift that publishers, marketers and users have already been experiencing. ChatGPT may still function as a chatbot, writing assistant and general-purpose AI interface, but when it searches the web, selects sources and synthesizes current information, it is also performing a search function at enormous scale.
That scale is what triggered the DSA designation. OpenAI reports approximately 159.1 million average monthly active recipients in the European Union for ChatGPT Search, far above the 45 million threshold used by the regulation.
ChatGPT Search has 159.1 million average monthly EU recipients
OpenAI’s Digital Services Act documentation provides an important qualification to the headline number. The 159.1 million figure refers specifically to ChatGPT Search—OpenAI’s online search features—and represents the average monthly active recipients in the EU over the six-month period ending March 31, 2026.
OpenAI explicitly says the average was calculated solely for compliance with its obligations under Article 24(2) of the DSA and should not be relied upon for other purposes. It should therefore not be treated as a general estimate of every European who uses ChatGPT each month.
For the regulatory question, however, the number is decisive. The DSA threshold for designation is 45 million average monthly EU users. ChatGPT Search is operating at more than three times that level.
The European Commission said ChatGPT, Reddit and Roblox had all declared that they reached at least the threshold required for designation, although Reddit and Roblox were classified as Very Large Online Platforms rather than search engines.
The EU is regulating what ChatGPT does, not what its interface looks like
The most interesting part of the decision is conceptual. A traditional search engine is easy to recognize: a user types a query and receives a ranked list of links. ChatGPT does not need to look like that to perform the same underlying discovery function.
Its online search features can retrieve current information from the web, identify relevant sources and construct a direct answer. The user may receive citations and links, but the primary experience is conversational rather than a conventional search-results page.
The Commission’s designation shows that the legal definition of search can accommodate that change in interface. A conversational AI product can simultaneously be a chatbot and an online search engine when its functionality meets the relevant criteria.
That does not mean every ChatGPT interaction is legally or technically identical to a Google query. Users can ask ChatGPT to rewrite text, analyze code, summarize uploaded files or brainstorm ideas without performing a public web search. The VLOSE designation reflects the significance and scale of its online search functionality.
OpenAI now has four months to meet additional DSA obligations
The designation is not merely a label. The Commission says the newly designated services have four months following notification—until January 2027—to comply with the additional obligations that apply to VLOPs and VLOSEs.
Those requirements are built around the idea that extremely large digital services can create risks that extend beyond individual users.
The Commission specifically says ChatGPT will have to assess and mitigate systemic risks stemming from its service and algorithmic systems. The areas named include dissemination of illegal content, negative effects on minors, users’ physical and mental well-being, fundamental rights, electoral processes and public security.
The broader DSA regime for very large services also brings stronger transparency and accountability requirements, including independent auditing and regulatory scrutiny of how systemic risks are identified and addressed.
For OpenAI, compliance therefore moves beyond responding to individual complaints. The company must demonstrate that it has processes for examining risks created by a search service used by more than 159 million average monthly recipients in the EU.
Generative search creates a different kind of information intermediary
The designation arrives as the distinction between search engines and answer engines becomes increasingly important.
A conventional search engine influences what people see by ranking sources. A generative search system can influence the information journey at additional stages: which sources it retrieves, which facts it selects, how it combines conflicting information and how it phrases the final response.
That can make the AI interface more useful, but it also concentrates more interpretive power inside the intermediary.
The user may never open the underlying websites. In some cases, the generated answer itself becomes the destination.
For regulators, that creates questions that go beyond the familiar problem of ranking transparency. Systemic-risk analysis may increasingly need to consider retrieval behavior, source selection, hallucination, synthesis and the way generated answers frame sensitive topics.
For publishers, ChatGPT is now a regulated discovery gatekeeper
The decision also matters for publishers whose content can appear as sources in AI-generated answers.
For decades, Google Search represented the dominant intermediary between much of the open web and its audience. Publishers optimized for rankings because those rankings could translate into clicks, advertising revenue, registrations and subscriptions.
AI search changes the transaction. A publisher can be cited by ChatGPT without necessarily receiving a visit, because the user may already have received the information needed from the generated answer.
That has produced disputes over crawling, copyright, attribution and whether AI answers create incremental discovery or substitute for traffic to original sources.
VLOSE status does not resolve those disputes, but it makes ChatGPT part of a regulatory framework explicitly designed for large information intermediaries. Publishers now have reason to monitor OpenAI’s DSA transparency and risk-management practices alongside its crawler policies, citation behavior and referral traffic.
For SEO, the decision confirms that AI visibility is a search problem
The search industry has spent years debating terminology for visibility inside ChatGPT, Gemini, Perplexity, Claude and other AI systems. GEO, AEO, AI SEO, LLM optimization and AI visibility all attempt to describe overlapping parts of the same transition.
The European Commission has now provided a formal regulatory signal: at least in the case of ChatGPT’s online search functionality, conversational AI belongs inside the search-engine category.
That does not make traditional SEO obsolete. It makes search visibility more fragmented.
A company can rank strongly in Google while appearing rarely in ChatGPT answers. Another company can be cited frequently by AI systems even when its own domain does not dominate conventional rankings. The sources that influence generated answers can also include third-party reviews, forums, news articles and reference sites.
Marketers therefore need to think beyond a single ranking system. If ChatGPT Search reaches 159.1 million average monthly recipients in the EU under the DSA measurement, its visibility cannot reasonably be dismissed as a niche experiment.
Transparency becomes harder when search produces one synthesized answer
The simplicity of conversational search hides considerable technical complexity.
A user asks one question and sees one answer. Behind it may sit query interpretation, web retrieval, ranking, source selection, model inference, safety systems and personalization.
Traditional search exposes at least part of that competition by displaying multiple results. Generative search can compress several sources into a single narrative.
This makes transparency especially important and especially difficult. Regulators need enough information to understand how systemic risks emerge without expecting companies to expose proprietary systems in ways that undermine security or competition.
OpenAI, meanwhile, must apply DSA compliance to a product whose outputs can vary according to the prompt, available sources, model version and context.
Independent audits will test processes, not individual answers
One misconception about regulation of generative AI is that compliance could mean verifying every answer before it reaches a user. At ChatGPT’s scale, that is not a realistic model.
The DSA’s systemic approach instead places emphasis on processes, controls and risk mitigation. Very large services are expected to identify significant risks created by their systems, implement reasonable measures and submit to independent scrutiny.
For an AI search engine, this could make questions about product design as important as individual content decisions. How does the service respond to high-risk queries? How does it handle illegal material? What protections exist for minors? How are election-related risks assessed? How are errors, appeals and transparency handled?
The Commission’s August 31 announcement makes clear that these are no longer merely voluntary trust-and-safety questions for ChatGPT in Europe. They are part of its regulatory obligations as a VLOSE.
The 159.1 million figure should not be turned into a generic ChatGPT audience statistic
Because the number is striking, it will inevitably be reused in marketing presentations and industry analysis. It needs context every time.
OpenAI says the figure is an average of monthly active recipients specifically for ChatGPT Search in the European Union, calculated over six months ending March 31, 2026. The company also says it was calculated solely for DSA compliance.
It is therefore evidence of regulatory scale, not a universal audience measurement comparable without qualification to web analytics, app users or conventional search-engine market share.
What it does establish is that ChatGPT’s online search features have crossed the EU’s threshold for a service considered systemically significant.
Search engines no longer need to look like search engines
The European Commission’s decision captures one of the biggest changes in online discovery since the rise of Google.
The underlying user need has not changed. People still want to find reliable information, compare products, understand events, research companies and answer questions. What is changing is the interface through which that need is satisfied.
Instead of receiving ten links and deciding which page to open, users can increasingly receive a synthesized response first and decide afterward whether any source deserves a click.
ChatGPT remains a chatbot and general-purpose AI assistant. But in Europe, its search functionality is now officially large and influential enough to be regulated as a Very Large Online Search Engine.
For OpenAI, that means a new tier of transparency, auditing and systemic-risk responsibility. For publishers and marketers, it is another sign that AI visibility belongs inside the search strategy rather than outside it.
And for the industry as a whole, the EU has formalized what user behavior was already making clear: in the age of generative AI, a search engine no longer has to look like Google.