US senators Maria Cantwell and Mike Rounds have urged White House AI czar Jay Clayton to develop a national strategy against AI-assisted cyber threats and include security standards for autonomous AI agents in his task force’s work. Their bipartisan letter follows a series of technology outages affecting US consumers and travellers, NBC News reported on 18 November. Cantwell is the top Democrat on the Senate Commerce Committee. Rounds, a Republican, chairs the Senate Armed Services Committee Subcommittee on Cybersecurity.
Key points
- The senators want a national strategy addressing cyber threats assisted by AI.
- They have asked Clayton’s task force to work with federal agencies and AI companies on digital infrastructure and standards for autonomous agents.
- US officials say they have found no evidence tying the recent outages to one another or to malicious actors.
- Clayton’s task force is due to produce an AI policy by the end of January.
Cantwell and Rounds seek agent standards
In the letter, the senators wrote that AI was increasing the speed at which weaknesses in software could be found and exploited, while defences were failing to keep up. They warned that AI could get around conventional cyber protections and that patches intended to protect infrastructure could take weeks or months to test and install. They also identified autonomous AI agents as a source of new security risks.
Their request is directed at Clayton’s task force, which brings together government officials and is charged with producing an AI policy by the end of January. Clayton was appointed to the White House AI role in early October and also serves as Director of National Intelligence. The senators want the task force to convene federal agencies and AI companies to strengthen digital infrastructure and develop security standards for agents that can act autonomously.
The letter asks for a strategy and for standards to be developed. They also cited incidents of rogue AI agents gaining access to customers’ bank accounts.
Japan’s banks enter the senators’ case
The senators cited a continuing series of cyberattacks in Japan affecting financial institutions and other companies. The attacks have led to large-scale breaches and leaks that the Japanese government is seeking to contain, and are believed to have been boosted by AI, NBC News reported. That account concerns attacks in Japan. The disruptions affecting US services are a separate part of the letter’s context.
Cantwell and Rounds also pointed to incidents in which rogue AI agents accessed customers’ bank accounts. According to their account, banks are worried that an agent could initiate transactions on its own. The senators asked for security standards for autonomous AI agents.
The senators cited AI-assisted discovery of software weaknesses and incidents in which agents accessed customers’ bank accounts. Their request brings AI companies into the proposed security work alongside federal agencies, rather than leaving the task force to develop its policy solely within government.
Rubio and Clayton address the US outages
Recent US disruptions have affected Microsoft Outlook, United Airlines and the Federal Aviation Administration. The FAA imposed a nationwide ground stop on 13 November amid instability in its radar and communications systems, Fox Business reported. Financial institutions and online services have also experienced outages. The senators’ letter arrived against that backdrop, although NBC News reported that the role of AI in the US outages was unclear.
Secretary of State Marco Rubio said on 17 November that officials had found no evidence connecting the outages to one another or indicating that hackers, criminals or foreign governments had caused them. Clayton also said officials had seen no indication of a coordinated cyberattack, and suggested system upgrades and security patches might explain the outages, Fox Business reported.
Ram Dantu, founding director of the University of North Texas’ Center for Cyber and AI Security, said the disruptions could be attributable to network congestion or software problems. He also said a cyberattack could not be ruled out and called for investigation. Dantu raised particular concern about the exposure of electricity, water and telecommunications systems to attacks.
Clayton is expected to testify before the House Committee on Homeland Security on 18 November, when questions about whether AI played a role in the outages are expected to arise.