On October 24, 2024, the White House launched a memorandum (the “Memo”) implementing Executive Order 14110 (“EO”), titled “Protected, Safe, and Reliable Artificial Intelligence,” which was issued in October 2023. The EO outlined a complete, all-of-government method to growing an AI governance framework. The Memo offers additional directives associated to AI governance, significantly within the nationwide safety context.
EO 14110 directed company motion in a spread of areas associated to AI, together with competitors and innovation, security and safety, client safety, staff’ points, privateness, fairness and civil rights, U.S. management overseas, and accountable authorities use of AI. The Memo builds on these themes and descriptions three primary aims [1] advancing the U.S.’s management in AI; [2] harnessing AI to satisfy nationwide safety aims; and [3] fostering the protection, safety, and trustworthiness of AI. The Memo directs a lot of companies, together with DOD, DHS, DOE, DOJ, CFIUS, NIST and others to realize the Memo’s aims. Highlights embrace the next:
- S. Management in AI: Make sure the U.S. stays the highest location for international AI expertise and computing services whereas defending U.S. AI from international intelligence threats.
- Promote progress, innovation and competitors: The Memo directs companies together with DOD, DHS, NSF and DOE to take actions comparable to streamlining the visa course of and fostering funding in AI infrastructure.
- Defend business, civil society, and educational AI mental property and infrastructure from international intelligence threats: The Memo directs the NSC and companies together with ODNI, DOD, DOJ, CFIUS and others to take actions comparable to reviewing the nationwide safety framework doctrine, figuring out vulnerabilities within the AI provide chain, and contemplating AI implications for lined transactions.
- Handle dangers. Proceed to develop worldwide AI governance with a spread of companions. The Memo directs NIST and Commerce (through NIST’s AI Security Institute) to take the lead in facilitating and offering steerage on AI testing packages and strategies.
- AI and National Safety: Responsibly harness AI’s energy to fulfill nationwide safety aims.
- Allow efficient and accountable use of AI. The Memo directs numerous companies to develop and entice AI expertise and encourage adoption of AI. The Memo additionally establishes a working group to handle AI procurement points, a directive with the potential to considerably influence the trajectory of AI throughout industries.
- Strengthen AI governance and threat administration. The Memo directs related companies to think about particular dangers instantly associated to every company’s use of AI and set up an AI Framework for the nationwide safety neighborhood.
- Protected, Safe, Reliable AI: Proceed to develop a “steady and accountable” framework for international AI governance. The Memo directs the State Division, in coordination with DOD, the Division of Commerce, DHS, the U.S. Mission to the UN, and USAID to develop a technique for partaking with international actors on AI governance in accordance with current frameworks.
The Memo additionally particulars the suitable use of AI in authorities, with a spread of directives associated to evaluation and reporting throughout companies with a categorized annex that covers different delicate nationwide safety points like countering adversary makes use of of AI. The Memo states that the current “paradigm shift” in AI towards giant language fashions and “computationally intensive techniques” primarily has occurred exterior of presidency up to now, however that it’s critically essential for the federal government to imagine a key function in AI governance and innovation.
With the primary of the Memo’s many directives due inside 30 days, and others following at numerous intervals, the U.S. authorities’s function in AI governance and threat administration will proceed to take form over the approaching yr.