ChatGPT is deactivating accounts without warning, even $200 Pro subscribers

ChatGPT users, including subscribers to the $200-a-month Pro plan, say OpenAI has been deactivating their accounts without warning since November 7. On login they see the error account_deactivated, and their chat history, saved memory and notes go with the account. Several say their appeals were rejected within minutes. The complaints were first collected by the Chinese tech site IT之家 (ITHome). As of November 8, OpenAI had not explained the wave or said how many accounts are affected.

What users report

On November 7 a Pro subscriber wrote on OpenAI's developer forum that their account had been deactivated that day with no warning and no notification email. According to ITHome, the user is enrolled in an OpenAI access program for cybersecurity specialists, and the login was protected by a physical security key. The user suspected the cause was connecting to ChatGPT and Codex through VPN servers in different regions. OpenAI's support account replied that it could not handle deactivation appeals on a public forum and pointed them to the link in the ban email or the appeal form in the help center. It gave no reason for the ban.

The same day another Pro user wrote on Reddit that their account had been closed for an alleged link to a previously banned account. The user said they had no other accounts, and the appeal was rejected about an hour later. Other users complained that notes and diaries they kept in ChatGPT were gone with the account.

On November 8 a user who said they had paid for ChatGPT for about four years posted on the forum that their first appeal was rejected automatically within minutes and a second, detailed one in about a minute. "That's too fast for anyone to have read it," they wrote. Every message to support since then, they said, gets an instant reply sending them back to the appeal, which is marked as closed. "Without any path to a human, it's a closed loop with no recourse."

What OpenAI's rules say

According to the help center as cited by Kod.ru and ITHome, an account can be deactivated for violating the usage policies, bypassing safety or access restrictions, sharing an account or API keys, suspicious activity or failing identity verification. A separate help article says that "accessing or offering access to ChatGPT or the OpenAI API from an unsupported country may result in your account being blocked or suspended". Russia, China and Hong Kong are not on the list of supported countries. Paying with a payment method from an unsupported country also gets users blocked.

A VPN by itself does not mean the rules were broken: someone in a supported country may simply use one for privacy. ITHome stresses that there is no proof the users were banned for this, or that all the cases belong to one enforcement action. Whether some of the bans are mistakes is also unknown.

Not the first wave

Kod.ru recalls a similar wave of forum complaints at the end of June, which also hit Pro subscribers. A week later one of the authors said their account had been restored with an apology for a wrongful ban, but the plan had been reset to the free tier. Claude users have complained too: according to Kod.ru, in October Anthropic banned accounts of users from Russia and Hong Kong who reached the service through a VPN, and called its decision on appeals "final".

How not to lose your data

A deactivated account takes everything stored in it, so a copy is worth making before anything happens. Free, Go, Plus and Pro users can request an export of their data in ChatGPT settings, and OpenAI warns that it is not available while logged out. The help center also offers a data request through OpenAI's Privacy Portal, but does not say whether it works for a deactivated account. Important notes and documents are safer kept outside the chat as well. If an account is blocked, the appeal goes through the link in the email from OpenAI or the help center form; the forum does not handle them. More on working with AI assistants safely is in our practical guide.