DOGE staffer with access to Americans' personal data leaked private xAI API key
2025-07-15
TechCrunch
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Marko Elez, a U.S. government employee, accidentally published a private API key granting access to 52 of xAI's large language models, including Grok, on GitHub. Although the repository was quickly removed, the key remained active, raising concerns about potential unauthorized use and broader AI security vulnerabilities.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event describes a security breach involving an AI system's API key, which could plausibly lead to unauthorized use or misuse of the AI models. Although the leak itself does not report any realized harm such as data breaches or misuse consequences, the potential for harm exists due to the exposure of sensitive access credentials. Therefore, this qualifies as an AI Hazard because it plausibly could lead to an AI Incident if exploited, but no direct harm has been reported yet.[AI generated]