How to detect hacks on ChatGPT, Claude, and Perplexity accounts
Original: How to tell if your AI platforms’ accounts have been hacked
Why This Matters
As AI platforms store sensitive conversation data, account security guidance is increasingly critical for mainstream users.
TechCrunch published a security guide on August 15, 2026 explaining how to detect unauthorized access on AI platforms including ChatGPT, Claude, and Perplexity. Each platform offers session management tools, though their authentication methods differ significantly.
TechCrunch published a practical security guide on August 15, 2026 detailing how users can check whether their accounts on major AI platforms have been compromised. The guide covers ChatGPT, Claude, and Perplexity — three of the most widely used AI services.
For ChatGPT, users can navigate to Settings > Security and Login > Active Sessions via the browser interface to view and terminate unrecognized sessions. Password resets are handled via a six-digit code sent to the user's registered email.
Claude, developed by Anthropic, uses a passwordless login system — sending a magic link to the user's email instead. Users can view and terminate active sessions under Settings > Account. Because there is no password, MFA is not supported.
Perplexity does not display active session locations. Instead, users can only perform a blanket 'Sign out of all sessions' action via the All Settings menu.
The article recommends using unique passwords stored in a password manager and enabling multi-factor authentication (MFA) where available. ChatGPT and Perplexity support MFA; Claude does not.