(Co-author: Matthew Tikhonovsky)
- On October 24, the Biden administration published the very first national security memorandum (NSM) on AI.
- In accordance with the executive decree of administration Biden on AI, the NSM describes concrete actions so that federal agencies take on that the United States government will open the way in the global development of AI, exploits the AI of its National Security Mission and promotes global consensus around AI.
- More importantly, the NSM directs the publication of the Framework to advance the governance of AI and risk management in national securitythat the White House has also published. Among other guidelines, the framework prohibits cases of using the AI with high impact by government agencies.
- While the Congress envisages legislation to establish similar requirements for the use of AI and by federal agencies, it is unlikely that such a legislation will adopt within five remaining weeks of this session.
On October 24, 2024, President Biden published the first National Security Memorandum (NSM) on AI. The NSM aims to galvanize “the adoption by the Federal Government of the AI to advance the National Security Mission”, while ensuring that such an adoption “reflects democratic values and protects human rights ». To achieve these objectives, the NSM describes various measures which must be taken by the government’s short -term and long -term government agencies.
The memorandum is under the Biden administration Executive Decree on AI (AI EO) and is one of the actions forced to be completed in the year following the signing of the order. Next week, we will face every one year actions that were to be announced before October 30, 2024.
The NSM focuses on three main objectives:
- To position the United States as a world development leader in a safe, safe and trustworthy AI
- To take advantage of AI cutting edge technologies to support the national security objectives of the US government
- To promote the international agreement and consensus on AI
1. To position the United States as a leader in the world development of a safe, safe and trustworthy AI.
- Flea development. The NSM aims to “improve the safety and diversity of flea supply chains” because “the development of advanced AI systems requires large volumes of advanced fleas”.
- Intelligence on competitors. The NSM makes the “collection on the operations of our competitors against our AI sector a high -level intelligence priority”. He also ordered relevant American government agencies to provide AI promoters with cybersecurity information in a timely and counterintelligence to protect their innovations.
- The IA security institute. The NSM designates the IA (AISI) Security Institute as the “main port of contact in American industry in the American government”. The AISI should facilitate “pre -public voluntary deployment tests for the safety, security and reliability of frontier AI models”.
- Resources for various AI developers. The memorandum orders the national research resource on AI to “distribute IT resources, data and other critical assets for the development of AI to a diversified range of actors who otherwise lack access to such capacities – such as universities, non -profit organizations and independent researchers. ”
- Evaluation of competitive advantage. The National Economic Council is responsible for coordinating “an economic assessment of the relative competitive advantage of the ECOC ecosystem in the United States private sector”.
2. To take advantage of advanced AI technologies to support the national security objectives of the American government.
- AI governance and risk management. The NSM “provides the very first advice for AI governance and risk management for use in national security missions”, focused on the maintenance of human rights and civil rights and to keep pace the rapid rate of technological change.
- Framework to advance governance and risk management in national security. The NSM calls for the establishment of a Framework to advance the governance of AI and risk management in national security (the frame), which was released with this NSM. The framework offers risk management, assessments, responsibility and transparency mechanisms. It also tasks in the government to identify and prohibit that “high impact AI uses cases according to the risks they represent to national security, international standards” and civil rights. It explicitly prohibits agencies from using AI to assign emotions, assess reliability or deduce the race.
- Rationalized supply practices. The NSM also orders various agencies to offer “rationalized supply practices and means to facilitate collaboration with non -traditional suppliers”.
3. To promote international agreement and consensus on AI.
- Strategy to advance AI governance standards. The NSM orders in the State Department, in coordination with other agencies, to “produce a strategy for the progress of international standards of governance of AI in accordance with the values of safe, secure and reliable and democratic values “, As well as commitment to international competitors and organizations.
- AI national security coordination group. The NSM also establishes the AI national security coordination group with the main officers of the AI of the State Department, the Ministry of Defense, the Ministry of Justice and a number of others Agencies. The group should focus on “the means to harmonize policies relating to development, accreditation, acquisition, use and evaluation of AI”.
Legislation to codify the requirements of use cases of high impact AI
The Congress is currently considering bipartite legislation which would co -order requirements, including those contained in the NSM and the framework, for the use of AI by federal agencies. As We have coveredThe law prepared for AI, introduced in June 2024 by senators Gary Peters (D-Mi) and Thom Tillis (R-N-NC), would establish a system of classification of the risk of AI and prohibit the use of AA by federal agencies to assign emotions, assess reliability, or infer the race.
However, as We have covered Last week, Congress has very little time to adopt the Act on the preparation of AI, which has been inactive since July. As We have already coveredWhile Senator Schumer (D-NY) indicated that he was planning to integrate AI bills into the essential end-of-year legislation, his main objective was to build regulations for elections generated by Deepfakes in such legislation. This orientation also reduces the probability that Congress will adopt laws concerning the use of the AI of this congress.