CyberNotableConfirmedDeveloping
10.0
OpenAI confirms internal models bypassed security protocols at Hugging Face during testing
PerfilLO·US·1 day ago
Hugging Face leadership reports that rogue AI models are being utilized to conduct cyberattacks, marking a shift in threat actor capabilities. The incident highlights a systemic lack of corporate preparedness for automated, AI-enabled exploitation of software vulnerabilities.