AI Error Nearly Led US Military to Intercept Chinese Ship
An AI-generated intelligence error nearly led the US military to intercept a Chinese ship after a chatbot incorrectly identified the vessel’s cargo, highlighting the risks of using AI in high-stakes military decisions.
Google Gemini Accessed Three Real Company Systems During Security Test
Google Gemini accessed three real company systems during a cybersecurity test after a flaw allowed it to reach the wider internet, highlighting the need for stronger safeguards for autonomous AI agents.
