Is someone else logged into your ChatGPT, Claude, or Perplexity account?
AI accounts can be broken into just like any other online account. Here is how to check for uninvited sessions on the three most popular AI platforms, and how to kick intruders out.
Key points
- ChatGPT, Claude, and Perplexity all let users check for active sessions, meaning logged-in devices they did not authorise.
- ChatGPT and Perplexity offer multi-factor authentication, an extra login step beyond a password; Claude does not use passwords at all.
- Perplexity is the only one of the three that does not show you which specific devices are logged in.
- Security experts consistently recommend a unique password for every account, stored in a password manager, an app that remembers strong passwords for you.
Your ChatGPT account holds your conversation history, any documents you have uploaded, and possibly a paid subscription. Someone who breaks in can read all of it. The same is true for Claude and Perplexity. These platforms are targets for exactly that reason.
First reported by TechCrunch, the good news is that all three services give you a way to find and remove suspicious logins. The steps are quick, and you do not need to be technical to follow them.
How do you check for uninvited logins on each platform?
Each service handles this slightly differently. Here is what to do on each one.
ChatGPT: Open ChatGPT in a desktop browser. Click your username in the bottom-left corner, choose Settings, then Security and Login, then Active Sessions. You will see every device currently logged in. Click the X next to any you do not recognise, or hit "Log out all" to clear every session at once. To change your password afterwards, log out, click Log In, enter your email, then choose "Forgot password" and follow the steps. ChatGPT will email you a six-digit code to confirm the change.
Claude: Open Claude in a desktop browser. Click your username, choose Settings, then Account. You will see Active Sessions listed there. Hover over any session you do not recognise, click the three-dot menu that appears, and choose Log Out or Terminate. There is no password to change because Claude never uses passwords. Instead, it emails you a login link each time. Remove a bad session and the intruder is locked out immediately.
Perplexity: Perplexity, an AI-powered search engine that answers questions by searching the web in real time, does not show you individual devices. If you suspect a problem, click your username, go to All Settings, and choose "Sign out of all sessions." That logs out every device, including yours, so you will need to log back in with your email address and a fresh six-digit code sent to your inbox.
What should you do right now to stay safer?
Three quick steps cover most of the risk. Use a different password on every account, store those passwords in a password manager (apps like these remember complex passwords so you do not have to), and turn on multi-factor authentication, or MFA, wherever it is offered. MFA adds a second check to the login process, usually a code sent to your phone, so a stolen password alone is not enough to get in.
ChatGPT and Perplexity both offer MFA. Claude skips it because the email-link system already acts as a built-in second check.
What should you watch for?
Four warning signs suggest your account may be compromised: unfamiliar session locations or device names in your active sessions list; AI responses that reference conversations you never had; a sudden change to your subscription or billing details; and login emails you did not request arriving in your inbox. If you spot any of these, follow the steps above to clear all sessions, then change your password immediately.
