How to check if your AI platform accounts have been hacked
ChatGPT, Claude and Perplexity each let users review active login sessions and sign out of unrecognized devices. TechCrunch has outlined the exact steps for checking and securing accounts on all three platforms.
Like any other online service, AI platforms such as ChatGPT, Claude and Perplexity can be targeted by hackers. Fortunately, all three offer ways to check whether an unfamiliar device is logged into your account.
Security experts generally recommend using unique passwords stored in a password manager and enabling multi-factor authentication (MFA), so a stolen password alone isn't enough to access an account. ChatGPT and Perplexity support MFA, but Claude does not, since Anthropic's chatbot uses email login links instead of passwords.
ChatGPT
To review sessions, open ChatGPT in a browser, click your username in the bottom-left corner, go to "Settings," then "Security and Login," and select "Active Sessions." Unrecognized devices can be logged out individually, or all sessions can be ended at once via "Log out all." To change a password, users must first log out, then select "Forgot password" on the login screen, enter their email, and use the six-digit code sent to set a new password.
Claude
In Claude, active sessions appear under "Settings" and then "Account." Hovering over an unfamiliar session reveals a three-dot menu with "Log out" or "Terminate" options, or all devices can be logged out at once. Since Claude has no passwords, re-entry always happens through a login link sent by email.
Perplexity
Perplexity does not display where an account is logged in. Instead, users worried about unauthorized access should go to "All settings," choose "Sign out of all sessions," and confirm. Logging back in requires an email address, after which a six-digit code or a direct sign-in link is sent.
