EU Designates ChatGPT a Very Large Online Search Engine

Introduction

On August 31, 2026, the European Commission designated ChatGPT as a Very Large Online Search Engine (VLOSE) under the Digital Services Act (DSA), and designated Reddit and Roblox as Very Large Online Platforms (VLOPs). The Commission's press release states that the three services "declared that they reach at least 45 million average monthly users in the EU and thus meet the threshold for designation." The Commission's public supervision register lists OpenAI Ireland Limited as the provider, ChatGPT as the designated service, and 159.1 million average monthly active users in the EU.

The headcount is not the interesting part. The classification is. The Commission wrote that ChatGPT "is an Artificial Intelligence (AI) system that can engage with and respond to users' prompts and queries, including by searching the web. Hence, ChatGPT is a hybrid service that qualifies as an online search engine under the DSA." A conversational assistant has been placed in the same DSA category as a crawled web index, and the register's enforcement column for ChatGPT currently holds one line: "31.08.2026: designation (decision not yet available)."

What the Designation Attaches

A four-month clock. The press release is explicit on timing: "Following the notification of the designations, these services have four months, i.e. by January 2027, to comply with the additional DSA obligations for VLOPs and VLOSEs." The Commission did not publish a specific day in January.

Systemic risk duties over the model, not just the content. The obligations cover "assessing and mitigating the systemic risks stemming from their service and algorithmic systems related to the dissemination of illegal content, the negative effects on minors, users' physical and mental well-being, fundamental rights, electoral processes and public security." The phrase "algorithmic systems" is what pulls a generative model's own behaviour into scope rather than only the third-party material it surfaces.

Standing compliance machinery. The Commission's VLOP/VLOSE policy page lists the recurring duties that attach on designation: an internal compliance function, an independent audit at least once a year with adoption of measures responding to the auditor's recommendations, data sharing with the Commission and national authorities, vetted-researcher access to service data where the research addresses systemic risks in the EU, a recommender-system option not based on user profiling, and a publicly available advertisement repository.

Investigative powers. Per the release, the Commission "will gain investigative powers to assess the functionalities behind these services and, where applicable, any related system."

Shared supervision. The press release states that the Commission supervises compliance jointly with Coimisiún na Meán, Ireland's Digital Services Coordinator, for ChatGPT and Reddit, and with the Authority for Consumers and Markets (ACM) in the Netherlands for Roblox. The Commission's own register does not match it on Reddit: it lists the provider as Reddit Netherlands B.V. with the Netherlands as its Digital Services Coordinator. The two Commission documents disagree on which authority co-supervises Reddit; they agree that Ireland co-supervises ChatGPT.

The three services designated on the day, with the user figures carried in the Commission's register:

Usability Analysis

For everyday users inside the EU, nothing changes today. The compliance obligations run against OpenAI, not against the product surface, and the four-month window means any visible consequences arrive later. What is worth watching is which of the VLOSE duties translate cleanly onto a chat interface and which do not.

Some map directly. An internal compliance function, an annual independent audit, and structured risk reporting are organisational requirements that any provider can stand up regardless of product shape. Vetted-researcher data access is potentially the most consequential of the set: on the Commission's own register it is the first time that route attaches to a generative answer engine, giving outside researchers a defined legal path to service data about systemic risks rather than relying on scraping or voluntary partnerships.

Others were drafted with feeds and ad auctions in mind. The recommender-system opt-out assumes a rankable, profile-driven stream of items; a model that composes an answer per prompt does not obviously have one. The advertisement repository requirement only bites to the extent the designated service actually carries advertising. Neither the press release nor the register explains how these are meant to apply to a hybrid service, and that translation work is what the next four months will consist of.

The register also puts the designation in scale context. Of the 28 services now under the DSA's strictest tier, only three are typed as search engines: Google Search (364 million declared EU monthly users), Bing (119 million), and now ChatGPT (159.1 million). Euronews described the classification as a first for an AI chatbot. An OpenAI spokesperson told PYMNTS that "ChatGPT search operates as a search service under the DSA, and we are preparing to meet the additional compliance requirements that come with this Very Large Online Search Engine designation."

Pros and Cons

The designation gives the Commission a supervisory handle over a consumer AI product that was previously covered only by the DSA's general obligations, and it does so through an existing, tested framework rather than waiting on new AI-specific machinery. It also creates a published, comparable user-number disclosure for a company that discloses relatively little.

Against that, the fit is imperfect. Reading a generative assistant as a search engine is a workable regulatory move but a lossy technical description, and several of the resulting duties were written for a product architecture ChatGPT does not have. The Commission has not yet published the designation decision itself, so the precise scope, in particular whether the classification reaches the whole assistant or only its web-search behaviour, is not yet on the public record.

Outlook

The near-term milestones are procedural: publication of the designation decision, then the first systemic-risk assessment and independent audit cycle ahead of the January 2027 deadline. Euronews notes that DSA non-compliance can draw penalties of up to 6% of global annual turnover, and the Commission's policy page states that a designation is revoked if a service stays below the 45-million threshold for a full year, which is not a realistic prospect here.

The broader signal is the precedent. The Commission's reasoning turned on a functional test, that the service "can engage with and respond to users' prompts and queries, including by searching the web," rather than on anything specific to OpenAI. Virkkunen's statement in the release closes with the line that the Commission "will not hesitate to designate any platform that meets the threshold for enhanced supervision under the Digital Services Act." Any assistant with web retrieval and EU scale now has a documented path into the same tier.

Conclusion

This is a jurisdictional decision rather than a product story, and its value is in the reasoning: the EU has answered the question of what a retrieval-capable chatbot is under existing platform law, and the answer is a search engine. For teams building EU-facing AI assistants, the practical takeaway is that DSA user-number disclosure and the 45-million threshold are now live planning inputs, and that the systemic-risk language reaches algorithmic systems, not only hosted content. The concrete obligations, and any enforcement, remain ahead.

Editor's Verdict

EU Designates ChatGPT a Very Large Online Search Engine brings real, demonstrable value, though with caveats that deserve weighing.

The strongest case for paying attention: applies an existing, tested supervisory framework to a consumer AI product instead of waiting for new AI-specific enforcement machinery. That alone raises the bar for what readers should expect in this space. Reinforcing that, forces a published, comparable EU user-number disclosure from a company that discloses relatively little operating data — practical value rather than just headline appeal. The broader signal worth registering is straightforward: the Commission's test was functional, not vendor-specific: a service that "can engage with and respond to users' prompts and queries, including by searching the web" qualifies as an online search engine, which sets a template any retrieval-capable assistant at EU scale can fall under. On the other side of the ledger, one constraint is real rather than a marketing footnote: classifying a generative assistant as a search engine is workable law but a lossy technical description of what the product does. It should factor into any serious decision. Layered on top of that, the non-profiling recommender option and advertisement repository duties assume a product architecture a per-prompt assistant does not have, and the Commission did not explain the mapping — which narrows the set of teams for whom this is an obvious yes.

For ChatGPT power users, OpenAI API customers, and enterprise teams already running on the OpenAI stack, a measured trial makes sense, with clear criteria for when to expand or pull back. For everyone else, the safer posture is to monitor coverage and revisit once the use cases that matter to your team are demonstrated in the wild.