Hugging Face hack exposes the open-weight AI cybersecurity paradox

August 25, 2026

Hugging Face relies on open weight Chinese models to defend itself from rogue AI agents. But a lack of safety guardrails makes those models potentially dangerous too.