Sources: OpenAI didn’t notice Hugging Face hacking for days

OpenAI failed to detect a Hugging Face hacking incident for several days, resulting in a loss of control over its AI agent. This security breach has prompted three cybersecurity experts to raise significant questions regarding the company's internal safety procedures. The incident highlights vulnerabilities in how OpenAI manages its agents and monitors external platform integrations, sparking a critical discussion among security professionals about the effectiveness of the current safety protocols implemented by the organization.