How to tell if your AI platforms’ accounts have been hacked
The article explains how users can check for unauthorised access to accounts on major AI platforms, including ChatGPT, Claude and Perplexity. It matters because these services may contain sensitive conversations or personal information, and compromised accounts can be misused even if the original password has been stolen.
Users are advised to use unique passwords held in a password manager and enable multi-factor authentication where available; ChatGPT and Perplexity support MFA, while Claude uses emailed sign-in links instead of passwords. ChatGPT and Claude show active sessions so users can remove unfamiliar devices or sign out everywhere, whereas Perplexity does not display sessions and instead requires users concerned about compromise to sign out of all sessions before logging in again.
- Check active sessions for unfamiliar devices.
- Sign out everywhere if an account may be compromised.
- Use unique passwords and multi-factor authentication.
New here? Start with this
AI platforms are online services that let people ask questions, create text and carry out other tasks using artificial intelligence. People often use them for work, study or personal matters, so their accounts may hold private conversations, uploaded files and details linked to an email address or payment method.
The main services include ChatGPT, made by OpenAI; Claude, made by Anthropic; and Perplexity, which combines AI-generated answers with web search. Access is usually protected by an account sign-in, but different services use different methods, such as passwords, emailed sign-in links and extra verification codes.
An account can be compromised if somebody else obtains access, for example through a reused or exposed password. Checking sign-in activity and protecting the account matters because an intruder could read information in it, alter settings or use the service while appearing to be the account holder.