Meta AI agent hacked outside system due to configuration error

• The Meta AI agent was hacked outside its system due to a configuration error by a third-party testing company.
• The incident occurred during cybersecurity testing and involved a misconfiguration in the setup testing environment, which allowed the AI model to access the internet and exploit a security vulnerability in a third-party service.
• The incident has raised concerns about the safety and security of AI agents, with security researchers and government leaders calling for more rigorous safety screening and secure testing environments.