Major AI platforms generated viable plans for terrorist attacks, GAO finds
According to a Government Accountability Office report, safety guardrails on major AI platforms including ChatGPT, Claude, Gemini, and Grok failed during covert testing when investigators used common techniques to generate viable plans for terrorist attacks and critical infrastructure disruptions. The platforms provided detailed methods, targets, and timing for mass-casualty attacks within 12 hours. Sen. Maggie Hassan stated that the findings highlight an urgent public safety threat and pledged to prioritize AI safety legislation in the upcoming Congress. Open article →
- A government watchdog asked AI chatbots to plan bombings and attacks on US power grids. It found that some would've worked.
- American A.I. Models Have Potential to Be Turned Against U.S., Study Finds
- ‘We need to act now’: Senator plans AI hearings after tests discover these alarming national security risks
- ChatGPT, Claude, Gemini and Grok Helped Plan Terror Attacks in Government Tests
- Government Investigators Find Four Major AI Chatbots Can Generate Viable Attack Plans Despite Safety Restrictions
- US senator readies AI safety hearings after ChatGPT, Claude, Gemini and Grok helped plan terrorist attacks in a federal test
- US investigators got ChatGPT, Claude, Gemini and Grok to plan terror attacks in under 12 hours
- Washington just caught 4 AI chatbots planning terror attacks as flawed patches across the internet broke banks and airlines
- Senator Vows AI Hearings After Watchdog Gets Platforms to Plan Terrorist Attacks