Senator vows AI hearings after watchdog gets platforms to plan terrorist attacks
The report says safeguards on ChatGPT, Claude, Gemini and Grok were bypassed, and each system produced attack plans in under 12 hours.
- On Wednesday, the Government Accountability Office released a report finding that four AI chatbots—ChatGPT, Claude, Gemini, and Grok—provided actionable plans for terrorist attacks after investigators bypassed initial safety guardrails.
- Democratic Sen. Maggie Hassan requested this May 2026 study to evaluate how commercial AI could exacerbate national security threats; researchers found all four platforms generated concrete playbooks for deadly attacks in less than 12 hours.
- Investigators focused on two attack types: improvised explosive devices and electrical grid disruption. Two models suggested vehicle-based IEDs, while others proposed drone attacks on substations, with all platforms suggesting locations within major United States cities to maximize damage.
- Responding to the report, Google, Anthropic, OpenAI, and SpaceXAI stated their systems have been strengthened since the spring, emphasizing commitment to safety. Hassan warned the AI public safety threat is "no longer theoretical."
- Congress will prioritize AI safety in upcoming sessions, though President Donald Trump stated Thursday that Washington opposes "any form of globalist regulation or anything to control or stifle" artificial intelligence, vowing the U.S. would not allow it.
153 Articles
153 Articles
Senate Committee to Hold January Hearings on AI Safety Risks Identified in GAO Audit
Senator Maggie Hassan, the incoming chair of the Senate Homeland Security Committee, has announced that the committee will hold hearings in January focused on AI safety and national security risks. The decision follows a performance audit by the Government Accountability Office (GAO), which concluded that four major AI platforms-ChatGPT, Claude, Gemini, and Grok-could be manipulated to generate viable plans for terrorist attacks within the Unite…
GAO: Four AI Chatbots Produced Viable Attack Plans in Tests
GAO says ChatGPT, Claude, Gemini and Grok each produced usable attack information within 12 hours of covert testing. The report's own caveats, and the companies' written objections, tell a narrower story than the headlines. The post GAO Covert Tests Got ChatGPT, Claude, Gemini and Grok to Produce Viable Attack Plans appeared on Right Side News.
US investigators got AI chatbots to plan mass-casualty, power grid attacks
ChatGPT, Claude, Gemini And Grok Produced Viable Terror Attack Plans ...
AI Chatbots Gave Up Attack Plans in Under 12 Hours, Watchdog Finds
AI Chatbots Gave Up Attack Plans in Under 12 Hours, Watchdog Finds A federal watchdog report says four leading AI chatbots, ChatGPT, Claude, Gemini and Grok, initially refused dangerous requests, but investigators used common jailbreaking tricks to get attack plans out of all four within 12 hours. Brian Baker and Johnette Cruz reacted on Baker […]
Coverage Details
Bias Distribution
- 55% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium

























![[your]NEWS](/_next/image?url=https%3A%2F%2Fgroundnews.b-cdn.net%2Finterests%2Ffb6dc495f74049f513563c33352175eaa0ecd509.jpg%3Fwidth%3D60&w=128&q=75)















