Is it safe to ask ChatGPT about immigration? Who sees your visa and status questions

The Morse teamPublished October 30, 20268 min read

It is safe to ask ChatGPT about immigration in general terms, like what an H-1B is or how a green card interview usually works, but a question about your own status, history, or case leaves a record you don't control. As of October 2026, OpenAI keeps that chat on its servers under your account, can be made to hand over account details on a subpoena and the chat itself on a warrant, and a border search can reach whatever the app keeps on your phone.

This isn't legal advice. It covers who can see an immigration question after you type it, not how to answer one. On October 30, 2026, we read OpenAI's, Anthropic's, and Google's privacy policies, government request policies, and latest transparency reports, the USCIS Policy Manual and its 2025 announcements, the State Department's June 2025 vetting notice, the federal statutes behind immigration subpoenas, a February 2026 court filing over a Homeland Security subpoena to Google, CBP's device search directive, and USCIS's and the FTC's warnings about notarios. Where something is only alleged, we say so.

Can USCIS or a visa officer see my ChatGPT chats?

Not through the online vetting the agencies have announced, which looks at public social media. USCIS said on April 9, 2025 that it would treat "antisemitic activity on social media" as a negative factor for people seeking green cards, foreign students, and others (USCIS). The Policy Manual, revised in an August 19, 2025 policy alert, now lists whether someone has "endorsed, promoted, supported, or otherwise espoused anti-American views" or a terrorist group's views "(including in social media content by, or involving an alien)" among the factors officers weigh, and calls those "overwhelmingly negative factors" (USCIS Policy Manual).

The forms follow the same line. A March 5, 2025 notice proposed collecting "social media identifiers ('handles') and associated social media platform names" on nine forms, including the N-400, I-485, and I-589 (Federal Register). For student and exchange visas, the State Department said on June 18, 2025 that it would vet "online presence" and instruct F, M, and J applicants to set "all of their social media profiles to 'public'" (State Department).

None of these documents mentions AI chats. A private conversation with ChatGPT isn't a profile anyone can set to public. It becomes public content only if you post it or share it by link. The routes to a private chat run through the company that stores it and the device that holds it.

Can ICE get my AI chat history from OpenAI?

It can get the account records behind a chat with a subpoena it writes itself, with no judge involved. Federal law gives "any immigration officer" the "power to require by subpoena" documents relating to "the privilege of any person to enter, reenter, reside in, or pass through the United States" (8 U.S.C. § 1225(d)). The Stored Communications Act lists what an "administrative subpoena" makes a provider produce: name, address, session times, "length of service," network addresses, and "means and source of payment" (18 U.S.C. § 2703(c)(2)). OpenAI, Anthropic, and Google all say a subpoena gets account records and the chats themselves take a warrant, as we lay out in can AI chats be subpoenaed.

So a chat about your status can be tied to your name, address, and payment card without anyone reading it. A February 2026 court filing shows Homeland Security using this subpoena on a tech company. A man's lawyers told a federal court in California that he emailed a Homeland Security attorney about an asylum case on October 30, 2025, and that about four hours later DHS issued an "Immigration Enforcement Subpoena" to Google for "a variety of private information" about him, his email account, and his use of Google's services. The filing alleges that two DHS agents and a police officer then questioned him at home about the email, and that DHS has a "pattern" of issuing such subpoenas and withdrawing them once challenged. It names four earlier cases in the same court that ended in dismissals. Homeland Security Investigations asked Google to retract this one on February 5, 2026, and the man dropped his motion five days later (notice of dismissal). That account is the filer's side, and no court ruled on it.

The AI companies' transparency reports count subpoenas and warrants but don't say which agencies sent them, so they can't show how often immigration agencies ask. We compare the reports in AI transparency reports.

Can a border officer read my AI chats?

Only what your phone or laptop holds. CBP's directive, in force since January 1, 2026, limits a border search to information "resident upon the device" and says officers "may not intentionally use the device to access information that is solely stored remotely" (CBP Directive 3340-049B). Two parts of it bear on immigration. Without probable cause, CBP may keep "only information" from the search "relating to immigration, customs, and other enforcement matters." And the directive doesn't cover information you choose to show, such as when "an alien proffers information to establish admissibility." What each AI app leaves on a phone is in what a border phone search reaches in your AI chats.

Are my AI chats about my immigration case privileged?

Not as of October 2026. In United States v. Heppner, which he called a question "of first impression nationwide," Judge Jed S. Rakoff held on February 17, 2026 that a defendant's chats with Claude were neither privileged nor work product, partly because Anthropic's privacy policy allowed disclosure to "third parties," including "governmental regulatory authorities" (memorandum). What that means for anyone asking AI about a legal matter is in private AI for legal questions.

Who could see an immigration question you type into an AI app?

We built this table from each company's and agency's own pages, read on October 30, 2026.

Google gives its users the plainest warning on the list: "Please don't enter confidential information that you wouldn't want a reviewer to see." Our guide to AI on a work laptop covers what each admin tool shows.

Can an AI chatbot give you immigration legal advice?

No. USCIS says "only an attorney or an accredited representative working for a Department of Justice recognized organization" can, and that "a notario público is not authorized to provide you with any legal services related to immigration" (USCIS, updated August 4, 2026). The FTC's guide, dated November 2023, adds that scammers call themselves "notario, notary public, accountant, or consultant," and warns against letting "a notario or anyone else keep your original documents" (FTC).

The FTC's warning about original documents carries over to chatbots. A passport scan, an A-Number, or a USCIS receipt notice pasted into an AI app is a copy of your papers held by a company that isn't your representative. Both agencies also say to check that any site claiming to be the government ends in .gov.

How Morse keeps your immigration chat history off company servers

Every route in the table except online vetting ends at a stored copy: a chat history on an AI company's servers, under an account tied to your email, phone, and card, or the same history cached on your phone. Morse is built without the server copy. Our servers don't keep messages or log what you write, and your Morse account holds your sign-in details, plan, balance, and a usage ledger with no content. Because we don't store chats, we usually have very little to give authorities (privacy policy).

Your history lives in your browser instead, encrypted and locked when you sign out. Unlocking it takes two keys, one that never leaves your device and one our server gives you when you sign in, so a copy of your browser data alone can't be read.

Each chat has its own privacy mode:

• E2EE, for questions about your own status, history, or case. Your browser checks the secure enclave's cryptographic proof before anything leaves your device, then encrypts the message so only that enclave can read it. Our servers relay text they can't read, and the reply is decrypted only in your browser. The enclaves are run by third-party providers, not by Morse, and the guarantee rests on the enclave hardware and attestation working as designed.
• Anonymous, to ask GPT, Claude, or Gemini how a visa category or form works. Our proxy removes your account, email, and IP address and sends the question under Morse's own account, so the provider sees it without knowing it's you. The provider may keep it for a limited time to check for abuse, under its own policy, so leave out names, A-Numbers, and dates.
• Private, for everyday questions. It's the default for a new chat. Open-weight models like DeepSeek, Kimi, and GLM, plus Grok, run under zero data retention agreements Morse has with its inference, search, and other providers, so they keep nothing.

In every mode, Morse doesn't train AI on your chats, sell your data, or show ads. A Temporary chat is kept in memory only and never saved, even on your device. Clearing your browser deletes your history for good, and Morse can't recover it, so keep anything your lawyer needs somewhere else.

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