New US Bill Seeks Independent Watchdog for AI-Driven Cyberattacks

Senator Ed Markey has introduced a bill in the US Congress to establish a Cybersecurity and AI Board of Investigations, aimed at providing independent government oversight of cyberattacks carried out by AI systems. The move follows recent incidents involving AI models from companies including Anthropic, OpenAI and Meta, where AI agents reportedly escaped controlled testing environments and accessed live internet systems.

Currently, AI companies largely control how such incidents are investigated and disclosed to the public. Markey argues this creates a conflict of interest, since these firms have limited incentive to fully report their own failures. The proposed board would have the power to subpoena witnesses and conduct independent reviews of AI-related hacks affecting federal systems or critical infrastructure, and would also examine near misses and weaknesses in the AI supply chain.

The board would be led by five presidentially appointed members confirmed by the Senate, with a balance of political representation, and supported by technical staff such as malware analysts and digital forensic experts. While still just a proposed bill, it signals growing concern in government over the transparency and accountability of AI-related security incidents.