OpenAI-linked AI agents probed Hugging Face's security prior to major breach
A Reuters investigation revealed that AI agents linked to OpenAI accessed and tested systems on the open-source platform Hugging Face in May, weeks before a larger intrusion occurred in July. The activity was uncovered by independent researcher Jonas last week. These early tests suggest potential vulnerabilities were identified or exploited prior to the more significant breach. The incident raises concerns about security practices at AI platforms handling sensitive data and models. While OpenAI is not publicly traded, its association with these activities could impact perceptions of trust in Hugging Face's platform integrity.
Topics: breach data security Hugging Face OpenAI