Home / Technology / AI Account Security: Spotting Intruders
AI Account Security: Spotting Intruders
15 Aug
Summary
- Check active sessions to identify unrecognized devices on AI accounts.
- Unique passwords and multi-factor authentication are key defenses.
- Claude uses email login links, lacking password-based MFA.
Users of popular AI platforms such as ChatGPT, Claude, and Perplexity should be aware that their accounts are susceptible to unauthorized access. To protect these accounts, it is recommended to use unique passwords, ideally managed through a password manager, and to enable multi-factor authentication (MFA) wherever available.
ChatGPT and Perplexity offer MFA, providing an extra layer of security. Claude, on the other hand, uses a different login method. It sends a link to your email address for authentication instead of relying on a password, meaning traditional MFA is not applicable. Checking active sessions in your account settings on ChatGPT and Perplexity can reveal unrecognized devices, allowing you to log them out.
If you suspect unauthorized access to your ChatGPT account, navigate to 'Settings,' then 'Security and Login,' and review 'Active Sessions.' For Claude, access 'Settings' and then 'Account' to view 'Active sessions.' Perplexity does not display active sessions, so users concerned about security should opt to sign out of all sessions via the settings menu.
To change your password on ChatGPT if compromised, you must first log out and then use the 'Forgot password' option. Claude does not have passwords to reset. After signing out of all sessions on Perplexity, you can log back in using your email, which will trigger a verification code sent to your inbox.
