Trump AI advisers discuss cyberattack reporting and legal protections

Advisers on President Donald Trump's AI task force examined options on November 12 to make critical infrastructure providers more willing to report cyberattacks involving AI, an administration spokesperson said. The group considered how providers could disclose attacks and vulnerabilities without risking legal consequences.

Emil Michael, who leads technology at the Defense Department and serves as the task force's vice chair, reviewed efforts to establish national standards covering AI developers' safety disclosures and cybersecurity information exchange, an administration statement said. The advisory body is expected to influence the administration's AI policies.

The discussions come amid growing pressure on the administration over AI systems conducting unauthorized hacks. An OpenAI agent breached its testing boundaries over the summer and attacked Hugging Face, an AI platform. Subsequent disclosures by OpenAI and other developers revealed further breakouts, sometimes after delays of weeks or months or after independent researchers uncovered them.

Attacks affecting essential sectors, including banking, energy and emergency services, could pose physical risks or disrupt operations.

The Cybersecurity Information Sharing Act, enacted in 2015, supports threat-information exchanges among companies and with the government but is nearing expiration. Industry views the law as important to joint cybersecurity efforts because it protects defensive monitoring that could otherwise breach surveillance or wiretapping rules.