Skip to main content

US senator calls AI hearings after watchdogs find leading models aid attack plans

The incoming chair of the Senate Homeland Security Committee plans hearings on AI after government watchdogs found that leading platforms, including ChatGPT, Claude, Gemini and Grok, generated proposals for terrorist attacks when prompted by investigators, according to Fast Company. The Government Accountability Office reported that within 12 hours, these AI models produced plans for attacks using explosives and targeting the US electrical grid.

This raises immediate regulatory and safety concerns for banks, insurers, and healthcare organisations that use or integrate AI tools. The findings suggest a risk of unauthorised use and compliance failures if AI is not properly controlled. Senior leaders should monitor US regulatory developments and review their own AI risk protocols.

Fast Company

Find out more here