In the waning months of the current administration, the White House issued a memo setting forth actions focused on national security as directed in the AI Executive Order from last year. As a reminder, the order -while directed to government agencies- also had impacts on how businesses use of artificial intelligence.
The national security memo builds on last year’s order, and of potential interest for businesses, calls on agencies including the Department of Defense, to assess potential risks to the US private sector’s AI competitive advantage. The memo also tasks the National Science Foundation with piloting programs to support AI development, and the Department of Energy and others to coordinate the support of AI-enabled infrastructure.
On the security front, the memo outlines the role for NIST as the government’s “primary point of contact” with the private sector’s AI developers. The memo tasks NIST with developing testing standards to ensure the safety, security and trustworthiness of AI models. Among other things, NIST is to look at the ability for AI models to impersonate people and ways to stop malicious use of AI models as well as risk management processes through the AI development lifecycle.
Putting it Into Practice: As the year draws to a close and we anticipate a new administration in Washington, we do not anticipate any slow down in the involvement of agencies on AI development, although the focus of some initiatives may change. The role of NIST as a conduit and developer of standards for the development of AI models will also likely continue.