DeepSeek vs ChatGPT privacy: the two apps side by side, with Morse alongside

The Morse teamPublished November 11, 2026Facts checked November 11, 202611 min read

On DeepSeek vs ChatGPT privacy, ChatGPT gives you more control and DeepSeek's app runs no ads, but both keep your chats on their own servers and both train on them by default. As of November 11, 2026, DeepSeek stores everything in China under Chinese law, keeps it as long as your account exists, and gives no deletion deadline. ChatGPT stores data in the US and other countries, deletes cleared chats within 30 days, and offers a temporary chat, but it shows ads on its Free and Go plans and lets reviewers read some chats. DeepSeek also publishes its model, and a copy run on someone else's servers never touches DeepSeek's app.

We read DeepSeek's privacy policy (last updated February 10, 2026), its terms of use (March 27, 2026), its training notes, and its help center on November 11, 2026. The same day we read OpenAI's privacy policy (updated September 10, 2026), its help pages on data controls, training, temporary chat, deleting chats, consumer data, and ads, its latest transparency report, and the regulators' own pages cited below. We ran no tests. In their favor, DeepSeek publishes its model weights for anyone to run, and OpenAI gives a day count for deleted chats, temporary chats, and ads data.

DeepSeek vs ChatGPT vs Morse at a glance

As of November 11, 2026. DeepSeek cells come from DeepSeek's pages, ChatGPT cells from OpenAI's, and the Morse column from our privacy policy.

Is DeepSeek safer than ChatGPT?

For privacy from the company, ChatGPT gives you more controls, and DeepSeek's app puts your chats to fewer uses. ChatGPT has a temporary chat, a 30-day deletion deadline, and an opt-out that leaves your history alone. OpenAI also publishes counts of the government requests it gets. DeepSeek offers none of those, but it runs no ads and sells nothing, and none of its pages describes staff or contractors reading chats the way OpenAI's do.

For privacy from governments, the deciding row is the law. DeepSeek's terms put every dispute under "the laws of the People's Republic of China in the mainland," heard by a court where Hangzhou DeepSeek is registered (DeepSeek). If that's your worry, the app is the wrong place for the chat, wherever you live. What DeepSeek's app collects about your device is in is DeepSeek safe.

Where do DeepSeek and ChatGPT store your data?

DeepSeek stores it in China, and OpenAI in the US and wherever its vendors are. DeepSeek's policy says plainly that "we directly collect, process and store your Personal Data in People's Republic of China," and that group companies handle it for "storage, content delivery, security, research and development, foundation model training and optimization" (DeepSeek). That wording has barely moved in three years, as our history of DeepSeek's privacy policy shows.

OpenAI is vaguer about place. Its consumer data page says content sits on OpenAI's systems and its "trusted service providers' systems in the US and around the world" (OpenAI), and the privacy policy adds that it uses "legally valid transfer mechanisms" when data crosses borders (OpenAI). It doesn't list the countries.

Do DeepSeek and ChatGPT train on your chats?

Both do by default, and both give the off switch the same name: "Improve the model for everyone." In ChatGPT it's under Settings, then Data controls. OpenAI says data sharing "is enabled by default" on personal Free, Plus, and Pro accounts, and that you can turn it off "and keep your chat history" (OpenAI). DeepSeek's terms, section 4.3, say it may use your inputs and outputs "to a minimal extent," after "strict de-identification," unless you turn the switch off (DeepSeek). Our guide to deleting a DeepSeek account shows where the switch sits on the web and in the app.

You don't have to email DeepSeek to opt out, whatever some guides say. AllAboutCookies' DeepSeek guide, updated September 28, 2026, still tells readers to "Email privacy@deepseek.com to make the request." DeepSeek's current terms, dated March 27, 2026, point to the switch instead, so you can flip it yourself.

The two opt-outs leak in different places. OpenAI lists its own gaps: a thumbs up or down can send "the entire conversation associated with that feedback" to training even after you opt out, and during Trusted Access sessions like Daybreak Blue, OpenAI may train on your content with the switch off (OpenAI). DeepSeek's training notes say the question-and-answer pairs it fine-tunes on come mostly from its research team, "with a small portion potentially based on user input" (DeepSeek). Its terms let you like or dislike a reply, but don't say whether a rated chat is treated differently.

How long do DeepSeek and ChatGPT keep your chats?

DeepSeek keeps them as long as your account is open, and ChatGPT until you delete them, plus up to 30 days. DeepSeek's policy keeps your account data, input, and payment data "for as long as you have an account," and longer for "a legitimate business interest" (DeepSeek). Its help center says "Deleted chats cannot be recovered" (DeepSeek), but none of its pages says when a deleted chat leaves its servers. Close the account and the terms still let DeepSeek "Retain certain data of the user as required by laws and regulations" (DeepSeek).

OpenAI gives a deadline. A chat you delete is removed from view at once and scheduled for permanent deletion "within 30 days," unless it was already de-identified or OpenAI must keep it "for security or legal obligations" (OpenAI). Files you saved to Library or Space stay until you delete them too. A temporary chat stays out of your history, makes no memories, and isn't used for training, though "OpenAI may keep a copy of a temporary chat for up to 30 days for safety purposes" (OpenAI). Neither DeepSeek's policy nor its terms describes anything like it. ChatGPT's other leftovers, memories among them, are covered in is ChatGPT private.

Who can read or demand your chats?

Some OpenAI staff and contractors can read them, and governments can demand them from both companies. OpenAI says authorized staff and "trusted service providers" may open your content for abuse or security investigations, support, legal matters, and model improvement unless you opt out (OpenAI). DeepSeek's terms give it the right to "review the behavior of users," with "risk filtering mechanisms" as the example, and its policy uses data to detect "abuse, fraud, and illegal activity" (DeepSeek). Neither DeepSeek page says whether a person ever reads a chat.

On governments, the policies sound alike and the records don't. DeepSeek may share data with "law enforcement agencies, public authorities, copyright holders, or other third parties" to comply with "applicable law, legal process or government requests," which it says it does "in accordance with the due process prescribed by law" (DeepSeek). OpenAI's policy lists six grounds, one of them a violation of its terms that OpenAI finds "in our sole discretion" (OpenAI). OpenAI also counts what it hands over. From January to June 2026 it received 146 content requests, disclosed data in 88, and got 250 to 499 US national security requests (OpenAI). We found no such report on DeepSeek's site or in a search on November 11, 2026. We compare the reports that do exist in AI transparency reports.

Does ChatGPT or DeepSeek use your chats for ads?

ChatGPT does on its Free and Go plans, and DeepSeek says it doesn't target ads at all. OpenAI began testing ads in the US on February 9, 2026. With ad personalization on, the ad you see below a reply can be picked using your current chat, "past chats and memory," and how you've handled earlier ads. Plus, Pro, Business, Enterprise, and Edu accounts see none, and neither do temporary chats (OpenAI).

Advertisers get only aggregated numbers such as views and clicks, and each ad setting is in how to turn off ChatGPT ads.

DeepSeek's policy says, in capitals, "WE DO NOT ENGAGE IN TARGETED ADVERTISING, 'SELL' PERSONAL DATA OR USE PERSONAL DATA FOR 'PROFILING.'" It does reserve the right to send you "commercial messages" (DeepSeek).

What have regulators done about DeepSeek and ChatGPT?

Italy's data protection authority has acted against both, and a Rome court has since sided with OpenAI against its ChatGPT fine. On January 30, 2025, the Garante ordered, "with immediate effect," a limit on how DeepSeek's two companies process Italian users' data, after it found their answers "del tutto insufficiente" (Garante). On December 20, 2024, it announced a €15 million fine and a six-month public information campaign for OpenAI, finding that OpenAI trained ChatGPT on personal data without first identifying a legal basis (Garante).

That fine was set aside in March 2026, when the Court of Rome upheld OpenAI's appeal, as we cover in AI chatbot privacy fines.

In the US, a Department of Homeland Security order published October 20, 2026 removes "all DeepSeek and DeepSeek-affiliate products, and third-party products and services developed using DeepSeek implementations" from federal civilian agencies' IT (Federal Register). It applies to agency systems, not to your own phone. South Korea's and Berlin's actions are in DeepSeek alternatives.

Does running DeepSeek's open weights change the comparison?

Yes, because DeepSeek's weights are free for anyone to run, which splits the model from the app, as we explain in is DeepSeek safe. Once the weights run on another company's servers, the China storage, the account-long retention, and the Chinese courts in the table above apply only to DeepSeek's own app. Whoever runs the model gets the prompt instead. ChatGPT's nearest equivalent is OpenAI's API, whose inputs and outputs OpenAI says it doesn't train on by default (OpenAI), but the prompt still goes to OpenAI.

The US order is the exception, since it names third-party products built on DeepSeek too. On a federal civilian agency's systems, a DeepSeek model on another host is still covered.

How Morse lets you use DeepSeek and GPT without either app's account

Picking DeepSeek's app or ChatGPT picks which company keeps a copy of your chat on its servers, under your account, and which country's law reaches that copy. Morse offers both companies' models in one account and doesn't keep your chats on its servers, so neither DeepSeek nor OpenAI holds your chat under your name.

• Private, for DeepSeek with no DeepSeek account. New chats start in Private, which runs open-weight models like DeepSeek, Kimi, and GLM. Morse has zero data retention agreements with its inference, search, and other providers, so they keep nothing. Where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't. There's no DeepSeek account keeping your history in China, and no "Improve the model for everyone" switch to find.
• E2EE, for questions you'd keep from every host, ours included. DeepSeek also runs in secure enclaves, which third-party providers operate, not Morse. Your browser checks the enclave's hardware attestation before anything leaves your device, then encrypts the message for that enclave alone, so our servers relay ciphertext they can't read. The guarantee rests on the enclave hardware and attestation working as designed, and E2EE has no web search or image generation. Here's how end-to-end encrypted AI chat works.
• Anonymous, for OpenAI's GPT models without an OpenAI account. Our proxy removes your account, email, and IP address and sends the message under Morse's own account, so OpenAI answers without knowing it's you. OpenAI may keep the message for a limited time to check for abuse, under its own policy, so leave out names and details that point to you.

Whichever model you pick, your history is kept in your browser, encrypted, and locked when you sign out. Morse doesn't train AI on your chats, sell your data, or show ads, so there's no ad picked from your past chats. Temporary chats are held in memory only, and since we don't store chats, we usually have very little to give authorities. Clearing your browser or deleting your account deletes your chats for good, and we can't recover them, so copy out anything you'll need.

Choose DeepSeek, ChatGPT, or Morse

• Choose DeepSeek if you want DeepSeek's newest build in its own app, with no ads and nothing sold, and you're comfortable with your chats stored in China for as long as your account exists.
• Choose ChatGPT if you want a temporary chat, a 30-day deletion deadline, published counts of government requests, and an opt-out that keeps your history, and you'll pay for Plus or Pro to drop the ads.
• Choose Morse if you want DeepSeek without a DeepSeek account, GPT without OpenAI knowing it's you, E2EE that your own browser checks, and history kept encrypted on your device, with no training or ads on any plan.

Send a first message from the Morse home page without an account, or start free with 15 messages a day and no card. Choosing your model, and E2EE chats, come with Pro, at $19 a month.

Morse publishes this comparison. We've tried to keep it fair; if something's wrong or out of date, email hello@trymorse.ai.

DeepSeek is a trademark of Hangzhou DeepSeek Artificial Intelligence Co., Ltd. Morse is not affiliated with or endorsed by Hangzhou DeepSeek Artificial Intelligence Co., Ltd.

ChatGPT is a trademark of OpenAI. Morse is not affiliated with or endorsed by OpenAI.