AI accounts now need the same incident-response habits as email, banking, and workplace messaging accounts. TechCrunch reports that accounts on popular AI platforms including ChatGPT, Claude, and Perplexity can be targeted by hackers, and it published platform-specific steps for checking whether an unfamiliar session may be active. The broad baseline is familiar: TechCrunch recommends unique passwords stored in a password manager and multi-factor authentication where the platform supports it. That second factor matters because a stolen password alone should not be enough to enter an account when MFA is enabled. The three services do not handle login security in the same way. According to TechCrunch, ChatGPT and Perplexity offer MFA. Claude does not offer the same password-based setup because Anthropic’s chatbot sends a login link to the user’s email address instead of asking for a password. For ChatGPT, TechCrunch says users can check active sessions from a computer browser by selecting the username in the lower-left corner, opening Settings, then Security and Login, then Active Sessions. That page shows where the account is logged in. If a device looks unfamiliar, the user can log out that specific device or choose to log out everywhere. Changing a ChatGPT password requires a logout first, according to the guide. From ChatGPT’s website, the user can choose Log in, enter the email address, select Forgot password, and continue through an emailed six-digit code before setting a new password. TechCrunch also notes that the password-reset email includes a link to official instructions. Claude’s check is similar but not identical. TechCrunch says users can open Claude in a computer browser, select the username in the lower-left corner, go to Settings, then Account, and review Active sessions. For an unfamiliar session, the user can hover over it, use the three-dot menu, and select Log out or Terminate; Claude also offers a log-out-of-all-devices option. The recovery mechanics differ because Claude does not use account passwords, according to TechCrunch. After ending sessions, the user logs back in with an email address and receives a login link. There is no Claude password to replace under that model. Perplexity provides less visibility into session details, TechCrunch reports. The service does not show where the account is logged in. If a user suspects access by someone else, the guide says to open Perplexity in a browser, select the username in the lower-left corner, go to All settings, and choose Sign out of all sessions, then Confirm. After that, Perplexity users can return through an email-based flow. TechCrunch says the user enters an email address and receives a six-digit code, or can use the Sign in button in the email to log in directly. Who benefits: Users who already rely on password managers, MFA, and routine session reviews have clearer steps to contain a suspected compromise. Teams standardizing on these tools can turn the platform-specific paths into internal checklists. Who's exposed: Users who reuse passwords or do not enable MFA where available face more risk from credential theft. Per TechCrunch’s guide, Perplexity users also have less session-location visibility because the service does not show where an account is logged in.