Is ChatGPT GDPR compliant? What OpenAI's EU pages say, and what regulators have found
The Morse teamPublished 11 min read
No regulator has declared ChatGPT GDPR compliant, and as of November 2026 no GDPR finding against it still stands. Italy's €15 million fine, the only GDPR fine on ChatGPT we found on a regulator's site, was annulled by the Court of Rome in March 2026. Ireland's Data Protection Commission, the authority now in charge of OpenAI in Europe, lists no decision on ChatGPT among its press releases. OpenAI has done the paperwork the law asks for: a European controller, a legal basis for each use of your data, and a contract for business customers. What's still contested is whether training on your chats under "legitimate interests" holds up, and whether ChatGPT can correct what it says about real people.
We read OpenAI's EU privacy policy (updated August 24, 2026), its help articles on training and data residency, and its enterprise privacy page on November 20, 2026. We also read the European Data Protection Board's ChatGPT taskforce report, a DPC record of a complaint against OpenAI Ireland, the Garante's release on its fine, noyb's two complaints, and the GDPR itself. OpenAI deserves credit for its EU policy. It sets out its legal basis for each use of your data in a table, names its Data Protection Officer, and tells you which regulator to complain to in the EEA, the UK, and Switzerland.
Morse, which publishes this page, is a private AI chat app with GPT, Claude, Gemini, DeepSeek, and uncensored models, encrypted so even Morse can't read your chats. If you're worried about a company holding your conversations and training on them, Morse stores your chat history in your browser, encrypted, not on Morse's servers, and no AI provider trains on what you send through Morse.
Is ChatGPT GDPR compliant on every plan?
The answer changes with the plan. On Free, Go, Plus, and Pro, OpenAI decides what happens to your chats as their controller. On Business, Enterprise, Edu, and the API, it processes them on your organization's behalf under a contract. OpenAI's EU privacy policy says it "does not apply to content that we process on behalf of customers of our business offerings, such as our API."
Read November 20, 2026, from the OpenAI pages listed under Sources:
| Plan | OpenAI's role for your chats | Data processing addendum | Where content is stored | Trains on chats by default |
|---|---|---|---|---|
| Free, Go, Plus, Pro (EEA, Switzerland) | Controller: OpenAI Ireland Limited | None for chats | Servers outside the EEA, including the US | Yes, until you switch it off |
| Free, Go, Plus, Pro (UK) | Controller: OpenAI OpCo, LLC, under the policy's "anywhere else" line | None for chats | Servers outside the UK, including the US | Yes, until you switch it off |
| ChatGPT Business | Processes it for your company | Yes | A region picked at checkout for a new workspace, plus a US copy "for a limited time" | No |
| ChatGPT Enterprise and Edu | Processes it for your organization | Enterprise: yes. Edu: a Student Data Privacy Agreement | One of 10 regions, Europe (EEA and Switzerland) among them, for new customers | No |
| API | Processes it for your company | Yes | A chosen region, for eligible customers | No |
Who is responsible for your ChatGPT data in Europe?
OpenAI Ireland Limited, if you live in the EEA or Switzerland. OpenAI's EU privacy policy names the Dublin company as "the controller" and adds that if you live "anywhere else," the controller is OpenAI OpCo, LLC in San Francisco. The same policy covers the UK, so taken at its word, a ChatGPT user in London deals with the US company as controller, under UK GDPR and the Information Commissioner's Office.
The Irish company matters because of the GDPR's one-stop shop. Article 56 makes the authority where a company has its "main establishment or ... single establishment" the lead authority for cross-border processing (EUR-Lex). The EDPB's ChatGPT taskforce report says OpenAI "was found to have a single establishment in the European Union since 15 February 2024" (EDPB). From that date, Ireland's DPC leads, and OpenAI's policy lists it "as our lead supervisory authority."
Not everyone accepts that Dublin decides anything. In its March 20, 2025 complaint to Norway's Datatilsynet, noyb named OpenAI OpCo, LLC as the respondent, called the Irish office "probably just a pro forma office," and argued that the real decisions about the model are made in the US (noyb complaint). That is noyb's argument, not a finding. On the other side, the Court of Rome upheld OpenAI's appeal against Italy's fine, and our page on AI chatbot privacy fines covers the court's reasoning on which regulator had the power to decide.
What legal basis does OpenAI use to train ChatGPT on your chats?
Legitimate interests. The legal-bases table in OpenAI's EU privacy policy says it trains on your content, when training is on or when you give feedback, "Where necessary for our legitimate interests and those of third parties and broader society." It relies on "perform a contract with you" to answer your prompts, and on "Your consent, for Free and Go users" to personalize ads.
That basis is the one Italy attacked. The Garante's December 20, 2024 release said OpenAI had trained ChatGPT on users' data "without first identifying an appropriate legal basis," failed to notify it of a March 2023 data breach, and had no age checks (Garante). That page now opens with a note that decision No. 755 "has been temporarily removed" after Court of Rome judgment No. 4153/2026, "published on 18 March 2026," which "upheld the appeal."
Because OpenAI relies on legitimate interests, the matching right is the right to object, which its EU policy lists alongside two training switches in your account; how to object to AI training under the GDPR walks through both.
Can ChatGPT correct false information about you?
OpenAI says it will try, within limits. Its EU privacy policy tells you "you should not rely on the factual accuracy of output from our models," and says that if you ask for a correction or removal, "we will consider your request based on applicable law and the technical capabilities of our models." Article 5(1)(d) of the GDPR requires personal data to be "accurate and, where necessary, kept up to date" (EUR-Lex).
A regulators' taskforce has already said a disclaimer isn't enough. In its preliminary views, the EDPB taskforce report says OpenAI's transparency measures help, but "they are not sufficient to comply with the data accuracy principle," and that "technical impossibility cannot be invoked to justify non-compliance" (EDPB).
noyb has tested this twice:
- Austria, April 29, 2024. noyb's complaint says a public figure asked OpenAI to fix the wrong date of birth ChatGPT kept giving, and that OpenAI refused to correct or erase it and answered that "factual accuracy in large language models remains an area of active research" (noyb).
- Norway, March 20, 2025. ChatGPT described Arve Hjalmar Holmen as a man convicted of murdering his children. noyb asked the Datatilsynet to order OpenAI to delete the output and "fine-tune its model to eliminate inaccurate results," and to impose a fine. noyb's page says ChatGPT now searches the web when asked about people and has stopped telling the story, but that the false data may remain in the training set (noyb).
Both are allegations in complaints, with no decision we could find as of November 20, 2026.
What has the Irish DPC done about ChatGPT?
The one DPC record on an OpenAI complaint we could find closed it as an amicable resolution, with no finding either way. It is dated May 9, 2025, and posted on the EDPB's website. A person in the Netherlands had asked OpenAI in June 2023 for every piece of their data used to train its models, and complained to the Dutch authority on October 4, 2023. The complaint reached the DPC on April 2, 2024 (DPC record).
OpenAI's answer, as the record describes it, shows the limit of a GDPR access request against a language model. OpenAI said its training data "is unstructured," that "there is no index, like that of search engine, allowing it to identify data related to specific individuals," and that it could search for a specific identifier such as an email address but couldn't find every mention of a person. It found no trace of the person's email in its training data and agreed to "filter out their verified personal data from future AI training model runs." The person said they "remained dissatisfied," the DPC heard nothing more from them, and the complaint was "deemed to have been withdrawn."
OpenAI can promise to leave you out of future training, but by its own account it can't show you everything a past model learned about you.
When did each GDPR action on ChatGPT happen?
Built from the primary documents above, read November 20, 2026:
| Date | What happened | Source |
|---|---|---|
| March 2023 | ChatGPT data breach, which the Garante's decision (since annulled) said wasn't reported to it | Garante release |
| April 13, 2023 | EDPB sets up a ChatGPT taskforce, since no one-stop shop applied yet | EDPB taskforce report |
| October 4, 2023 | Access complaint against OpenAI Ireland filed in the Netherlands | DPC record |
| February 15, 2024 | OpenAI's single EU establishment in Ireland recognized; the DPC becomes lead authority | EDPB taskforce report |
| April 29, 2024 | noyb's accuracy complaint filed in Austria | noyb |
| May 23, 2024 | EDPB taskforce report: warnings about output "not sufficient" for accuracy | EDPB taskforce report |
| November 2, 2024 | Garante decision No. 755: €15 million fine and a six-month information campaign | Garante release |
| December 20, 2024 | Garante announces the fine and says it has forwarded its file to the DPC | Garante release |
| March 20, 2025 | noyb's complaint in Norway, naming OpenAI OpCo, LLC | noyb complaint |
| May 9, 2025 | DPC records the Dutch complaint as amicably resolved | DPC record |
| March 18, 2026 | Court of Rome judgment No. 4153/2026 published, upholding OpenAI's appeal | Garante release |
| August 24, 2026 | Current version of OpenAI's EU privacy policy | OpenAI |
Does ChatGPT store your data in the EU?
Not on a personal plan. OpenAI's EU privacy policy says it "processes your Personal Data on servers located outside of the EEA, Switzerland and the UK," including in the United States, and relies on adequacy decisions and the European Commission's standard contractual clauses for the transfer.
Storage in Europe is something OpenAI sells to organizations, on ChatGPT Enterprise, Edu, Business, and the API, with different terms on each. We compare what each offer covers in what is data residency.
Can a business use ChatGPT under the GDPR?
On a business plan with a signed DPA, it can start to. OpenAI's enterprise privacy page (updated January 8, 2026) says it can "execute a Data Processing Addendum (DPA) with customers for their use of ChatGPT Business, ChatGPT Enterprise, and the API," and that Edu and Teachers fall under a Student Data Privacy Agreement. Its training article says OpenAI doesn't use Business, Enterprise, Edu, or API inputs and outputs to improve its models by default.
A personal plan is the problem. On Free, Plus, or Pro, OpenAI is the controller of whatever you paste in, and there's no DPA for chats, so a customer's or employee's details go to a company your own privacy notice may never mention. The data processing agreement explainer covers what the contract must say, and is ChatGPT Business private covers what your employer's admins can see. A DPA settles OpenAI's side; your company still needs its own legal basis for what it puts in. This isn't legal advice.
What can you do if you use ChatGPT in Europe?
Use the rights OpenAI's EU privacy policy lists: to access, delete, correct, move, and restrict your data, and to object to processing based on legitimate interests. In practice:
- Stop training. Switch off Improve the model for everyone, and skip the thumbs up and down on chats you'd rather keep out.
- Use Temporary Chat for one-off questions. The policy says those chats are deleted automatically "within 30 days," unless OpenAI must keep them for safety or legal reasons.
- See what OpenAI holds. Export from your data controls, or send a request through privacy.openai.com or dsar@openai.com. What does ChatGPT know about you walks through both.
- Ask for a correction. The same portal takes requests to fix or remove false output about you.
- Complain. Article 77 lets you complain to the authority where you live, work, or where the alleged breach happened (EUR-Lex). OpenAI's policy names the DPC as its lead authority, or your local authority, plus the ICO in the UK and Switzerland's FDPIC. Its Data Protection Officer is at dpo@openai.com.
Deleting has limits too. OpenAI's policy says deleted data is removed "within 30 days," except content that "has already been de-identified and disassociated from your account" for training.
How Morse stores your chat history in your browser
OpenAI keeps your conversations on its servers, as their controller, and decides what to train on. Every access request, correction request, objection, and data transfer above turns on that stored copy.
Morse is built so there's much less of it. Your chat history is stored in your browser, encrypted, not on Morse's servers, and Morse's servers keep no log of your messages. Morse doesn't train AI on your chats, sell your data, or show you ads, and no AI provider trains on what you send through Morse. Morse sends your message under its own account, so the model provider never learns your name, email, account, or IP address. You choose a mode for each chat:
- Anonymous, for GPT, Claude, or Gemini without your name attached. The provider reads the message without knowing it's you, and may keep it for a limited time to check for abuse, under its own policy, so leave out other people's names and details.
- Private, for chats you'd rather no model host kept. Morse has zero data retention agreements with its inference, search, and other providers, so they've agreed to keep nothing, and where a provider confirms zero retention on each reply, Morse checks it.
- E2EE, for what no one else should read. Your browser checks the enclave's hardware attestation, then encrypts the message so only that verified enclave can read it. Morse's servers relay text they can't read, and every E2EE reply has a lock you can open to see each check that passed.
Because Morse doesn't store chats, it usually has very little to give authorities. If you clear your browser, your chats go with it, so save a backup to a file from Settings > Data. The details are on the Morse privacy page.
Send a first message from the Morse home page without an account, or start free with no card. E2EE chats come with Pro, at $19 a month.
Sources
Read November 20, 2026.
- OpenAI: EU privacy policy, enterprise privacy, How your data is used to improve model performance, Data residency and inference residency for ChatGPT, Where your ChatGPT Business content is stored
- European Data Protection Board: Report of the work undertaken by the ChatGPT Taskforce, May 23, 2024
- Data Protection Commission (Ireland): press releases; record of amicable resolution, complaint concerning OpenAI Ireland Limited, May 9, 2025
- Garante per la protezione dei dati personali: release on the ChatGPT fine, with its note on the Court of Rome judgment
- noyb: Austrian complaint release, April 29, 2024; Norwegian complaint release and the complaint itself, March 20, 2025
- European Union: General Data Protection Regulation, Articles 5, 56, and 77
Common questions
Is ChatGPT banned in the EU?
No. OpenAI Ireland Limited offers it to people in the EEA under OpenAI's EU privacy policy. The only GDPR fine on ChatGPT we found, Italy's €15 million from 2024, was set aside by the Court of Rome in a judgment published March 18, 2026.
Who is the data controller for ChatGPT in Europe?
OpenAI Ireland Limited, in Dublin, if you live in the EEA or Switzerland. For anyone else, OpenAI's EU privacy policy names OpenAI OpCo, LLC in San Francisco, and read literally that includes people in the UK.
Is ChatGPT Enterprise GDPR compliant?
OpenAI signs a data processing addendum for ChatGPT Business, ChatGPT Enterprise, and the API, and offers storage in Europe to new Enterprise and Edu customers. A DPA covers OpenAI's side; your company still needs its own legal basis for whatever personal data it puts in.
Does ChatGPT store data in the EU?
Not for personal plans. OpenAI's EU privacy policy says it processes your data on servers outside the EEA, Switzerland, and the UK, including in the United States, under the EU's standard contractual clauses or adequacy decisions. EU storage is sold to organizations.
How do I send OpenAI a GDPR request?
Use the tools in your account first, then privacy.openai.com or dsar@openai.com, which OpenAI's EU privacy policy lists for access, deletion, correction, portability, restriction, and objection. Its Data Protection Officer is at dpo@openai.com.
Can OpenAI correct false things ChatGPT says about me?
You can ask, and OpenAI says it will consider the request 'based on applicable law and the technical capabilities of our models.' Two complaints by the privacy group noyb, in Austria in 2024 and Norway in 2025, argue that this falls short of the GDPR's accuracy rule.
Where do I complain about ChatGPT under the GDPR?
In the EEA, to Ireland's Data Protection Commission as OpenAI's lead authority, or to your own national authority. In the UK, to the Information Commissioner's Office, and in Switzerland, to the Federal Data Protection and Information Commissioner.
Does Morse keep my chats or train on them?
Your chat history is stored in your browser, encrypted, not on Morse's servers, and Morse's servers keep no log of your messages. Morse doesn't train AI on your chats, and no AI provider trains on what you send through Morse.