Meta says its AI model breached a third-party company during testing - CBS News

Meta's AI model accidentally breached a third-party company's systems during testing, raising concerns about AI security and unintended harms. This highlights real risks as AI systems become more powerful and widely used.