How to Check a Hacked AI Account on ChatGPT, Claude and Perplexity

AI accounts can contain conversation histories, uploaded files and other personal information, making unauthorised access worth taking seriously. ChatGPT, Claude and Perplexity provide different security controls for checking or ending suspicious sessions.

If an unfamiliar device or session appears, users should remove it promptly and secure the account. The exact steps differ because the three platforms use different login systems.

How to check your ChatGPT account

ChatGPT users can open Settings and review Active Sessions in the security section. The page shows sessions associated with the account, including available information about devices and sign-ins.

If a session looks unfamiliar, it can be ended individually. Users can also choose to log out of all active sessions when they suspect broader unauthorised access.

For accounts that use an OpenAI password, users can reset it on the ChatGPT login page by selecting the password recovery option and following the email verification instructions. Accounts created through another login provider may need to be secured through that provider instead.

ChatGPT also supports multi-factor authentication, which adds another verification step to the sign-in process.

How to check active sessions in Claude

Claude users can review active sessions from their account settings. Suspicious sessions can be terminated individually, and the service also provides an option to log out of all devices.

Claude uses passwordless authentication for users who sign in through email, sending a login link or verification method to the registered email address instead of relying on a conventional account password.

That makes the security of the connected email account particularly important. Anyone who suspects unauthorised access to Claude should terminate any unfamiliar sessions and ensure that access to the email account used for signing in is secure.

How Perplexity handles suspicious sessions

Perplexity takes a different approach. Rather than providing the same device-by-device session view as ChatGPT and Claude, users can go to account settings and choose to sign out of all sessions.

After signing out of all sessions, users can authenticate again using Perplexity’s supported login methods. Email-based login can involve a one-time code or sign-in link sent to the registered address.

What to do if an AI account may be compromised

Finding a session from a device or location that cannot be explained is grounds for revoking that access. Signing out everywhere provides a broader reset when it is unclear which session may be unauthorised.

Users should also secure the authentication method connected to the AI service. That can include changing an account password where applicable, reviewing the security of the associated email or third-party login account, and enabling multi-factor authentication when the platform supports it.

The most important difference among the three services is how much session information they expose. ChatGPT and Claude allow users to inspect active sessions, while Perplexity provides a broader sign-out control. Knowing where those settings are can make it easier to respond quickly when an account shows signs of unauthorised access.