What happens to your AI chats when you die: what 11 companies say

The Morse teamPublished October 27, 202611 min read

When you die, your AI chats usually stay on the company's servers, unread by your family, until an inactivity timer or a deletion request removes them. As of October 27, 2026, OpenAI, Anthropic, xAI, Character.AI, Replika, and Perplexity publish no process for a deceased user at all. Google, Apple, and Meta let you name someone ahead of time, but Meta's legacy contact can't read your messages, Apple's Legacy Contact gets none of your saved passwords, and Google's help page doesn't say whether Gemini chats are included.

We read each company's own terms, privacy policies, and help pages on deceased users, inactive accounts, and legacy contacts on October 27, 2026, and built the table below from them. Where a company publishes nothing, we say so. Estate law varies by country and state, so this isn't legal advice.

What happens to your AI chats when you die, company by company?

It depends on the company, and for most of them on a timer rather than a process. "Family" here means anyone other than you, including an executor.

The platforms built for photos, email, and posts publish deceased-user processes. The six companies whose main product is AI publish none, so their terms decide what happens to your chats, mostly through their inactivity clauses.

Can your family get your ChatGPT history after you die?

Not through any process OpenAI publishes. Its terms, effective January 1, 2026, say "you may not share your account credentials or make your account available to anyone else," and that you "may not assign or transfer any rights or obligations under these Terms." The privacy policy's route for someone acting for you asks the agent to "present authority to act on your behalf, such as signed written permission," sent to dsar@openai.com (OpenAI). Neither page mentions death. That route, and the access requests it serves, are in our guide to downloading your AI chat data.

The gap has already come up in court. In a wrongful-death lawsuit filed in December 2025, the plaintiffs alleged that "OpenAI won't produce the complete chat logs" of a ChatGPT user who had died, and Ars Technica reported that it had confirmed "OpenAI currently has no policy dictating what happens to a user's data after they die" (Ars Technica, December 15, 2025). Those are the plaintiffs' claims, not findings.

Claude is in the same position. Some estate-planning guides say Anthropic's terms forfeit a Claude account at death, but the consumer terms we read on October 27, 2026 don't mention death. They bar sharing your login, bar transferring the terms "without our prior written consent," and let Anthropic close a free account after a year of inactivity (Anthropic). Anthropic's privacy policy accepts requests from an authorized agent with "evidence of authorization to act on behalf of the data subject" (Anthropic).

How long until an inactive AI account is deleted?

Between 120 days and two years, where a company names a number. xAI's terms, updated October 12, 2026, let it "delete your account at any time without notice" if it "has been inactive for more than 120 days," and the clause names no exception for paid plans (xAI). OpenAI and Anthropic wait over a year and promise notice first, and only for accounts without a paid plan. Google and Microsoft close whole accounts after two years. Character.AI and Replika reserve the right to close accounts inactive "for an extended period of time" and give no number.

The chats can go before the account does. Gemini Apps Activity is auto-deleted after 18 months by default, so by the time Google's two-year clock runs out, chats kept on the default setting are already gone (Google). Microsoft says it stores Copilot "conversation activity for 18 months" by default (Microsoft). Snapchat runs the other way: "All the content you share with My AI is retained until you delete it" (Snapchat), and Snap's page for families offers deletion only by logging in to the account.

Does Google's Inactive Account Manager include Gemini chats?

Google doesn't say. Inactive Account Manager lets you pick up to 10 people to receive "all or only specific data types" once your account has been inactive for a time you choose, and notes that "some information can't be shared" (Google). The help page doesn't list the data types, which you see only while setting it up.

Gemini chats sit in your Google activity, and Google's own export steps reach them through My Activity, then Gemini Apps, in Google Takeout (Google). If you set up Inactive Account Manager, check whether My Activity or Gemini appears in the list, and tick it if you want it passed on. Without a plan, Google says it "can work with immediate family members and representatives to close the account" and "in certain circumstances" may provide content. Ask for the contents before asking for closure, because once you choose to close the account, Google "is unable to process any request to turn over the contents of the account at a later date" (Google). Neither route reaches chats a human reviewer read, which Google keeps apart from your account for up to three years, as we cover in is Gemini private. Our guide to deleting Gemini history covers the activity settings.

Can a legacy contact see your AI chats?

Not through Apple or Meta. An Apple Legacy Contact can get iCloud Photos, Notes, Mail, Messages in iCloud, iCloud Drive files, and anything in an iCloud Backup, but not "information stored in the account holder's iCloud Keychain, such as user names and passwords" (Apple, published April 2, 2026). So the saved passwords to your AI apps don't pass on. The contact needs the access key you created plus a death certificate, and the Legacy Contact account lasts "three years from when Apple approves" the first request before it's permanently deleted (Apple).

Meta's legacy contact can write a pinned post, change the profile photo, request removal, and download "a copy of what you've shared on Facebook," but can't "log into your account" or "read your messages" (Meta). A Meta AI chat isn't something you shared, and Meta's AI help pages publish no process for it. What Meta keeps from those chats is in is Meta AI private. If you'd rather it all went, Delete after death removes "all of your messages, photos, posts, comments, reactions and info" from Facebook once someone tells Meta you've died (Meta).

Microsoft has no contact to name. It says "You don't need to contact us to let us know that someone has died," closes the account after two years of inactivity, and considers releasing contents only after "a valid subpoena or court order" served on its registered agent, with exceptions for heirs in Germany and close relatives in China (Microsoft).

Does a will give your family access to your AI chats?

It can help, depending on where you live. The Uniform Law Commission, which drafts model laws for US states, wrote the Revised Fiduciary Access to Digital Assets Act. Minnesota's version, adopted in 2016, lets you allow or forbid disclosure of your digital assets, "including the content of electronic communications," in a will or other record, but a direction given through a company's online tool overrides a contrary will (Minnesota Statutes 521A.04). For the content of communications, the company must disclose to the estate's representative only if you consented or a court directs it, and the representative has to bring a death certificate, proof of appointment, and the record of your consent (521A.07).

None of the six AI companies without a deceased-user process offers an online tool of that kind, so for those chats, a will or another written record is where a direction like that would go. Whether a chat with an AI counts as an "electronic communication" under your state's law is a question for an estate lawyer.

How do you plan for your AI chats now?

Decide what each account should do, then act while you can still sign in:

1. Export what you want kept. Each app's export steps are in our guide to downloading your AI chat data. Store the files where the person you trust can reach them.
2. Delete what you don't. Deleting is faster than any timer, and each app has its own steps. ChatGPT's are in our guide to deleting ChatGPT history for good, and Character.AI's, including what happens to characters you published, are in how to delete a Character.AI account.
3. Use the tools that exist. Set up Google's Inactive Account Manager, an Apple Legacy Contact, and a Facebook legacy contact or Delete after death.
4. Write the rest down. List the AI apps you use and what you want done with each, and keep it with your will.

How Morse keeps your chats from waiting on a server

Every route above runs through a company's servers, where your chats sit under its terms until a timer, a court order, or its own discretion decides what happens to them. Morse never keeps them there. The history of every Morse chat lives in the browser you used, encrypted, and nowhere on our side. We don't keep your messages or log their content. Our records hold how you sign in, your plan, your balance, and a usage ledger that carries none of what you wrote.

So an executor has no Morse archive of your chats to ask for, and no 120-day or two-year clock is counting down on them. Signing out locks the history. Opening it again needs two keys together, one kept on your device and never sent anywhere, the other handed over by our server at sign-in, so someone who copies the browser's files gets nothing they can read. Because we don't store chats, we usually have very little to give authorities. We also don't train AI on what you write, sell your data, or run ads.

The same design means your history ends with that browser. Clearing it, losing the device, or deleting your account erases the chats, and we have no way to bring them back. Anything you want someone to have, keep somewhere they can reach: paste it into your own notes or documents, and record where they are alongside your will.

You also choose, chat by chat, who sees a message on its way to the model:

• E2EE, for things no one should ever read, during your life or after it. Before a word leaves your device, your browser verifies the enclave's hardware attestation and seals the message so only that enclave can open it. What passes through our servers is ciphertext we can't read. Third-party providers run the enclaves, not Morse, and the protection depends on the enclave hardware and attestation working as designed.
• Private, so no model host holds a copy that outlives you. New chats start here. It runs open-weight models such as DeepSeek, Kimi, and GLM, plus Grok, and Morse has zero data retention agreements with its inference, search, and other providers.
• Anonymous, for GPT, Claude, or Gemini without an account of yours at OpenAI, Anthropic, or Google. Your email, account, and IP address are stripped before the message goes out under Morse's own account, so the provider reads it without knowing it's you. It may keep the message for a limited time to check for abuse, under its own policy.

What should you do about each kind of AI chat?

• Chats you want your family to have: export them from each app and store the files with your estate papers. On Morse, copy what matters into your own notes, since the history lives only in your browser.
• Chats you want gone: delete them in each app now, and choose Delete after death on Facebook. On Morse, there's no copy on Morse's servers to delete.
• Chats you don't want decided by a company's timer: keep them out of server-side history to begin with. Morse stores them only on your device, encrypted, and our privacy policy has the details.

Try a chat from the Morse home page without an account, or start free with no card. End-to-end encrypted chats come with Pro, at $19 a month.