How to Check If Your ChatGPT, Claude or Perplexity Account Was Hacked

Summary

TechCrunch cybersecurity reporter Lorenzo Franceschi-Bicchierai’s August 15, 2026 guide says OpenAI’s ChatGPT, Anthropic’s Claude and Perplexity accounts can be breached like other online services. He recommends unique passwords stored in a password manager and multi-factor authentication. ChatGPT and Perplexity support MFA, while Claude uses passwordless emailed sign-in links. All three provide session removal, but their visibility and recovery controls differ.

On ChatGPT’s computer browser interface, Settings, Security and Login, then Active Sessions shows logged-in devices and permits individual or universal logout. Changing the password requires logging out, selecting Log in, Forgot password and Continue, then entering an emailed six-digit code and a new password, or following the email’s reset link. Claude’s Settings and Account page lists Active sessions; users can open an unknown session’s three-dot menu to Log out or Terminate, or log out every device. Reentry uses an emailed link because Claude has no passwords. Perplexity shows neither logged-in devices nor locations, so users must open All settings, choose Sign out of all sessions and confirm. They can return with an emailed unique six-digit code or the email’s Sign in button.