The EU Now Regulates ChatGPT Like a Very Large Search Engine

The European Commission has designated ChatGPT as a Very Large Online Search Engine under the Digital Services Act. The label, usually shortened to VLOSE, follows ChatGPT’s declaration that it reaches at least 45 million average monthly users in the European Union. OpenAI now has four months to meet additional obligations for the service.

This is a scale-triggered regulatory status, not a review of one model benchmark. The Commission is treating ChatGPT’s search-like public information function as large enough to create systemic effects that require more formal controls.

Why ChatGPT is classified as a search engine

ChatGPT began as a conversational assistant, but users increasingly ask it to find, summarize and synthesize information. That behavior overlaps with online search even when the interface does not look like a list of blue links.

The DSA distinguishes very large platforms from very large search engines. Reddit and Roblox were designated as VLOPs in the same announcement, while ChatGPT received the VLOSE classification.

The legal category focuses on the service’s function and scale in the EU. It does not mean ChatGPT is technically identical to Google Search or Bing. It means the Commission considers the search function within the DSA’s scope.

For publishers and product teams, the classification is another sign that answer engines are part of information distribution. Optimizing only for conventional result pages misses where users now seek explanations.

What changes over the next four months

The Commission says designated services must assess and mitigate systemic risks stemming from their services and algorithmic systems. The listed areas include illegal content, effects on minors, physical and mental well-being, fundamental rights, electoral processes and public security.

Very large services also face stronger transparency, data-access and audit expectations. The exact implementation belongs in compliance documents and future Commission supervision, not in a one-paragraph press release.

The deadline is four months after notification. Official EU pages differ in cached language about whether that lands at the end of November or December, so a compliance team should use the formal notification date and legal counsel rather than a media countdown.

The important point is that the designation starts a compliance window. It is not a statement that all additional controls were already complete on August 31.

What users may notice

Many obligations operate behind the interface: risk assessments, audits, researcher access and internal governance. Users may eventually see clearer explanation of recommendation or ranking behavior, reporting tools and transparency documentation.

The DSA also includes rights around platform decisions and advertising transparency for covered services. How those requirements map to a conversational search engine will be one of the practical questions.

Do not expect every answer to gain a regulatory badge. The law does not turn a chat response into a government-approved statement. Users still need to evaluate sources, recency and uncertainty.

Our ChatGPT tier comparison addresses product access and pricing. The DSA designation is a separate layer about service governance in the EU.

Search quality and safety are not the same metric

A model can produce a fluent, relevant answer and still create a systemic risk through repeated bias, unsafe recommendations or opaque source selection. Conversely, a highly cautious model can frustrate users without necessarily satisfying every transparency duty.

Regulation therefore needs evidence at several levels. Individual answer accuracy matters. Aggregate patterns matter. Appeals, logging, audit access and incident response matter.

OpenAI will need to explain how it identifies and mitigates risk without disclosing information that would make abuse easier or expose user privacy. That tradeoff is familiar across security programs.

The designation may also push clearer separation between generated synthesis and source evidence. Users benefit when a claim can be traced, especially for current events or high-stakes topics.

What this means for businesses using ChatGPT

The VLOSE designation does not replace a company’s own obligations. A business using ChatGPT to draft hiring decisions, medical communications or financial advice still owns its workflow and sector-specific rules.

Procurement teams should ask which service and region they use, how data is handled, what logging exists and how users can report failures. A consumer account, an enterprise product and an API integration are different contracts and data paths.

Companies should also keep independent records for consequential work. A chat history is not automatically a compliant case file.

For organizations, the safest interpretation is operational: track which product surface is used, preserve source evidence, and route consequential outputs through the same review controls that would apply if a human researcher supplied them.

Our ChatGPT, Claude and Gemini comparison discusses capability differences. Regulatory status adds another comparison dimension, but it does not create one universal winner.

What publishers should watch

If ChatGPT functions as a large search engine, publisher concerns about attribution, crawling, source visibility and traffic become part of the distribution debate. The DSA designation does not automatically resolve licensing or compensation.

Researchers may gain structured ways to study systemic effects under the DSA. That could improve evidence about how answer engines represent sources and topics, provided privacy and security controls are sound.

Publishers should maintain clear canonical pages, dates, authorship and source evidence. Those signals help human readers and automated systems understand the content, even though no publisher can control how a model answers every query.

The worst response would be to generate more low-value pages simply because answer engines exist. Scale regulation increases the case for distinctive, sourced material rather than bulk repetition.

What happens next

OpenAI must move through the compliance window, while the Commission will supervise the designated service with Ireland’s digital-services authorities playing a role. Formal decisions, risk reports and audits will provide better evidence than speculation about interface changes.

Legal challenges or implementation disputes are possible. The designation itself is nonetheless clear: ChatGPT’s EU reach and search function have crossed the DSA’s very-large-service threshold.

For users, the practical rule remains the same tonight: a popular answer is not automatically correct. For the provider, popularity now carries a more formal duty to measure and mitigate risks at scale.

The Commission’s designation also offers a useful baseline for future comparisons. ChatGPT is listed separately from Reddit and Roblox because the legal service category differs, even though all crossed the scale threshold. Compliance reporting should preserve that distinction rather than importing every social-platform duty into an AI-search headline without checking the statute and formal decision.

Users outside the EU may see product changes, but the designation directly governs the covered European service. Global rollout should be confirmed, not assumed.

Regional terms and notices remain decisive.

That distinction prevents false compliance claims.

FAQ

What is a VLOSE? A Very Large Online Search Engine under the EU Digital Services Act.

Why was ChatGPT designated? It declared at least 45 million average monthly users in the EU and provides a search-like service.

Is ChatGPT banned in Europe? No. The designation creates additional compliance duties.

When must OpenAI comply? The Commission gives designated services four months after notification; teams should use the formal notice for the exact deadline.