Your AI Says No, Then Does It Anyway
New research proves AI models will refuse harmful requests verbally while executing them through tool calls
Tag
Tag
New research proves AI models will refuse harmful requests verbally while executing them through tool calls