ChatGPT Ad Transparency Is Legally Due in January 2027

ChatGPT ad transparency has a hard deadline now, not a vague one. The European Commission designated ChatGPT a Very Large Online Search Engine (VLOSE) under the Digital Services Act on August 31, 2026, the first AI chatbot to get that label.

The trigger was blunt: roughly 159.1 million average monthly EU users, more than three times the 45 million threshold. Four months later, in January 2027, OpenAI owes the EU an ad repository. Nobody has built one for a conversational interface before, and no platform bound by the same rule has shipped a complete one either.

TL;DR: ChatGPT ad transparency has a legal deadline: OpenAI must publish an Article 39 ad repository (full text) by January 2027, four months after its VLOSE designation under the Digital Services Act. Every platform already bound by that rule (X, TikTok, Meta, Google) ships an incomplete one, so plan for a lagged, broad-category archive rather than a Meta Ad Library equivalent.

Why ChatGPT ad transparency is now a legal question

Designation isn't a warning letter. It's a status change with a clock attached. The European Commission's August 31, 2026 decision named ChatGPT alongside Reddit and Roblox as newly designated Very Large Online Platforms and Search Engines, and ChatGPT's case is the odd one: it's the first conversational AI product to cross the VLOSE threshold.

The number that did it was reach, not ad spend or complaint volume. OpenAI reported roughly 159.1 million average monthly users across the EU, well past the Digital Services Act's 45 million threshold for VLOP/VLOSE status. Cross that line and a separate, heavier compliance regime switches on automatically. Size triggers it, not conduct.

That matters for the ad transparency question specifically because Article 39, the DSA's ad repository requirement, only binds designated VLOPs and VLOSEs that serve ads. ChatGPT started serving ads in 2026. The designation is what turns that ad product into a legal obligation.

The ad business grew fast enough to make the deadline consequential. OpenAI piloted ads to US users on the Free and ChatGPT Go tiers on February 9, 2026, opened a self-serve Ads Manager with $3-5 CPC bidding that May, and expanded to 31 European markets on August 24, 2026 — the rollout that put ad-serving inside the jurisdiction Article 39 covers.

OpenAI's ad revenue reportedly reached a $1 billion annualized run rate in under 200 days. When we tested ChatGPT ads for two weeks, that pace matched what showed up in-market before the European rollout even started.

The January 2027 deadline and what else is due

Designated platforms get four months to comply. For ChatGPT ad transparency specifically, that lands the compliance deadline in January 2027, and the ad repository is one obligation among several arriving at once.

The fuller bundle includes systemic-risk assessments (OpenAI has to document how ChatGPT could amplify illegal content, disinformation, or harm to minors), an independent third-party audit of its DSA compliance, and a researcher data-access channel so vetted academics can query platform data directly. Ad transparency is the most visible piece of that bundle because it's the one meant for the public, not regulators or researchers.

As of this writing, no repository exists. OpenAI hasn't published a beta, a technical spec, or even a placeholder page. That's normal at this stage of a four-month runway, but it also means every claim about what the repository will contain is a forecast, not a report — a distinction this piece keeps separate throughout.

What Article 39 actually forces platforms to publish

The Article's text is specific, which is unusual for EU platform law. For every ad shown, a compliant repository has to carry seven fields:

• The ad's content, including the product, service, or brand advertised
• Who the ad was presented on behalf of
• Who paid for it, if that's a different party
• The period during which it ran
• Whether it was targeted, and the main targeting parameters, including any exclusion criteria
• A flag marking it as a commercial communication rather than organic content
• Total recipients reached, plus aggregate breakdowns per EU member state

Removed ads don't disappear from the record. They stay in the repository minus the advertiser's identity, which is the Article's one concession to takedown requests. (This section describes the legal text, not advice on how to act on it — nothing here is legal advice.)

Read that list again with a marketer's eye: it's a spec for a static placement. A start date, an end date, one piece of creative, one advertiser, one targeting configuration. That assumption is going to matter later in this piece. Developers already pulling data from live repositories under this rule have documented what's actually accessible in a DSA ad repository today, and it's a useful preview of the gap between the statute and the export.

The compliance record other platforms have already set

Article 39 isn't new, and the wider ad transparency data landscape already has years of history to check before trusting any prediction about ChatGPT ad transparency on day one. X, TikTok, Meta and Google have all been bound by it, against the standards our own breakdown of ad transparency libraries and regulatory standards covers in more depth. Social Media Transparency scored each platform against the Article's eleven sub-requirements, and the pattern is consistent: nobody's clean.

Across every examined platform, the same five gaps repeat: payer identity, targeting mechanics, searchability by ad content, API access and data currency. Four sets of engineers, four legal teams, four sets of regulators leaning on them for years. The failure modes converge anyway, and that's the ad compliance base rate a reader should apply to January 2027, not optimism about a clean-slate build.

TikTok is the case worth watching closest. It's the platform closest to full compliance, with a hard December 2026 fix deadline a month before ChatGPT's own. It's also the only one of the four with a working non-commercial API. Google's own search interface (our guide to the Ads Transparency Center covers its search mechanics) is the furthest from useful for anyone tracking competitor spend by hand. If OpenAI copies a homework assignment, TikTok's is the one still being graded.

Why a chat interface breaks Article 39's assumptions

Here's where the piece turns from fact to analysis, and it's worth being explicit about that switch: everything from here to the FAQ is reasoning from the pattern above, not a report on what OpenAI has committed to build.

Article 39 was written for a feed. An advertiser buys a placement, it runs from a start date to an end date, one creative asset serves every viewer who matches the targeting, and "the period during which it was presented" is a clean pair of timestamps.

A generated response has none of that. Ask ChatGPT the same question twice and an ad-eligible reply can differ each time: different phrasing, potentially different product framing, no fixed creative asset to point a repository row at.

"The content of the advertisement" gets slippery too. Is the unit of record the sponsor relationship, or every distinct generated variant of the response that mentioned it? The Article doesn't say, because nobody drafting it in 2022 was picturing a chat window.

Regulators building on Meta's ad model may end up applying a placement-shaped rule to a product that has no placements, and the gap between the two is where this repository is most likely to look thin on day one.

What ChatGPT ad transparency will probably still miss

Treat this section as a forecast built from the compliance record above, not as anything OpenAI has confirmed. No EC delegated act or technical-format guidance for conversational-AI ad repositories turned up in this research, and it's worth saying plainly: it may not exist yet.

Given that every platform already bound by Article 39 has landed on the same five failure points, the reasoned expectation for ChatGPT's repository is:

• Not searchable by ad content, following Meta and Google's pattern rather than a purpose-built exception
• Lagged rather than real-time, closer to Google's roughly 90-day cadence than to a live feed
• Targeting disclosed as broad targeting categories, not the demographic targeting, behavioral targeting, or contextual targeting specifics Article 39(e) technically requires
• Total reach reported in aggregate, not a per-ad impression count a researcher could audit
• Advertiser identity that's self-declared and unverified, the same weak point auditors flagged in Meta's repository

None of that is a knock on OpenAI specifically. It's what four other engineering and legal teams converged on under the same statute, with more runway than OpenAI has.

Agencies already working through how to prepare for ChatGPT ads shouldn't wait for January 2027 to build a research habit. Competitor ad research that depends on a single regulatory filing, published once a quarter in broad categories, is a weaker habit than one built on data queried today. The same holds for teams building a cross-platform ad strategy that has to cover ChatGPT alongside Meta and TikTok.

Some research teams are already wiring ChatGPT to live ad data through custom GPT actions rather than waiting on a filing at all. Separately, adlibrary's unified ad search indexes live creative across Meta, Google and TikTok in one interface. Its AI ad enrichment tags advertiser and product fields structurally instead of leaving them to self-declaration, and its ad timeline analysis tracks when a creative entered and left rotation, the exact "period during which it was presented" field Article 39 struggles to define for a chat product.

None of that replaces a regulator's repository once ChatGPT's ships. It's a parallel data source, not a substitute for one. It won't cover ChatGPT itself until OpenAI opens something to query. The bigger shift underneath all of this, how ChatGPT ads will change attribution before they change media buying, will outlast whatever format the repository ships in.

FAQ

Is ChatGPT part of the Digital Services Act? Yes. The European Commission designated ChatGPT a Very Large Online Search Engine on August 31, 2026, which brings it under the DSA's full compliance regime, including Article 39's ad transparency rule.

What is Article 39 of the Digital Services Act? Article 39 requires designated platforms that serve ads to publish a public repository covering ad content, the advertiser, the payer, the display period, targeting parameters, a commercial-communication flag, and reach totals per EU member state.

When does ChatGPT have to publish ad transparency data? ChatGPT ad transparency data is due by January 2027, four months after its August 31, 2026 VLOSE designation, alongside systemic-risk assessments, an independent audit and researcher data access (advertising obligations under the DSA).

Does the Meta Ad Library show who paid for an ad? Partially. Meta shows an advertiser name, but that identity is self-declared and unverified, one of the gaps regulators have flagged across multiple Article 39 examinations.

What is a Very Large Online Search Engine under the DSA? A search engine or platform with more than 45 million average monthly users in the EU, a threshold that triggers heavier obligations including risk assessments, audits and ad transparency.

Compliance under Article 39 has been a multi-year project for platforms with far more lead time than OpenAI has now, and none of them finished it. ChatGPT ad transparency will arrive in January 2027 as a repository, not a Meta Ad Library equivalent built for a chat interface. The safer plan is to expect a lagged, broad-category archive that answers fewer questions than it raises, and to treat anything more complete as a pleasant surprise rather than the design intent.

Want a competitive research habit that doesn't wait on a regulatory filing? Start a free trial of adlibrary, or check pricing.