TechCrunch outlines steps to secure AI platform accounts against unauthorised access
Users are advised to check active sessions and enable multi-factor authentication to mitigate risks of compromised accounts on major generative AI services.
TechCrunch has published a comprehensive guide detailing how users can detect and mitigate unauthorised access to accounts on major artificial intelligence platforms, specifically ChatGPT, Claude, and Perplexity. The publication highlights that, like other online services, these AI interfaces are targets for hackers seeking to compromise user credentials.
The guide recommends fundamental cybersecurity practices, including the use of unique passwords stored in a password manager and the activation of multi-factor authentication (MFA) where supported. This second layer of security is critical to prevent access even if primary credentials are stolen.
For ChatGPT users, the platform provides visibility into active login locations. Users can navigate to Settings, then Security and Login, to view Active Sessions. If an unfamiliar device is identified, it can be logged out individually or all sessions can be terminated simultaneously. Password resets are conducted via email verification codes.
Claude, developed by Anthropic, utilises a different authentication model that relies on email-based login links rather than traditional passwords. Consequently, the platform does not offer MFA. Users can manage access by viewing Active Sessions under the Account settings and terminating suspicious logins. Because there is no password to change, securing the account involves revoking existing sessions and re-authenticating via email.
Perplexity does not display specific login locations, requiring a different approach to account hygiene. Users concerned about unauthorised access are advised to go to All settings and select the option to sign out of all sessions. Re-entry requires verification through a six-digit code sent to the registered email address.
The publication notes that interface elements and navigation paths may evolve as platforms update their software. Users are encouraged to review these protocols regularly to ensure their accounts remain secure against emerging threats.
