Out of 33 non-public names tested, Grok gave accurate residential addresses in multiple cases and rarely refused a request. It often returned dossiers of personal data no one asked for. Other major chatbots (ChatGPT, Gemini, Claude) all declined to do the same due to privacy protections.
2/3
xAI’s own policies say violating people’s privacy is prohibited. Grok is doing it anyway, and xAI didn’t respond to Futurism’s questions.
This is one of the nightmare scenarios privacy advocates have been warning about around AI being used for nefarious actions.
3/3
If you have a fediverse account, you can quote this note from your own instance. Search https://mastodon.social/users/oaklandprivacy/statuses/115662262881390586 on your instance and quote it. (Note that quoting is not supported in Mastodon.)