Secure, Governable Chips
The Center for a New American Security (CNAS), in collaboration with the Institute for AI Policy and Strategy, has released a new report, Secure, Governable Chips, by Onni Aarne, Tim Fist, and Caleb Withers.
The report introduces the concept of “on-chip governance,” detailing how security features on AI chips could help mitigate national security risks from the development of broadly capable dual-use AI systems, while protecting user privacy. The report also offers a set of recommendations for U.S. policymakers:
The White House should establish a NIST-led interagency group via executive order to coordinate on-chip governance policy.
The Commerce Department should create commercial incentives for chip firms to improve AI chip security using advance market commitments.
NIST should coordinate with government funding bodies to scope, fund, and support crucial AI chip security R&D in academic and private labs.
The Departments of Commerce and Homeland Security should establish an AI chip red-teaming program and flexible export licensing policies to support a staged rollout of on-chip governance mechanisms.
The Departments of State and Commerce should coordinate with key allies in the AI hardware supply chain to develop shared policies and standards for on-chip governance.
Leading chip firms should move early to build and harden the additional security features required for robust on-chip governance.