Maintaining American Leadership in Artificial Intelligence (Executive Order)
On February 11, 2019, the Executive Office of the President published Executive Order 13859, “Maintaining American Leadership in Artificial Intelligence.” In the ten sections of this Artificial Intelligence (AI) focused Executive Order, the President outlines his “American AI Initiative”, and associated policies, principles, and objectives for sustaining and enhancing what it identifies as the United States’ scientific, technological, and economic leadership in Artificial Intelligence research, development, and deployment. Also included in the Executive Order are the President’s consideration of Federal data and computing standards, regulation of artificial intelligence, its impacts on the American Workforce, and protection of Federal and private intellectual property.
In section one, the Executive Order outlines the President’s belief in the emerging and dominating significance of Artificial Intelligence (AI), the technology’s Research and Development (R&D), and the government’s role in ensuring the technology’s use to “enhance our economic and national security, and improve our quality of life.” Specifically, this Executive Order identifies the President’s five guiding principles and motivation for its creation.
American AI Initiative guiding principles:
- Public/Private Leadership: Federal, academic, and private investment to maintain the country’s leadership in AI R&D;
- Standard Setting and Reducing Barriers to Innovation: setting the appropriate safety standards and deployment of AI for its adoption in Industry and new AI-enabled industries.
- Training for AI Now and its Future: training workers to ensure economic competitiveness for the current and future economies;
- Public Trust: fostering trust and confidence in the operation of AI as well the protection of civil liberties and safety in its operation; and
- International Openness and Competitiveness: promoting an international environment of fair and competitive markets where AI is used.
Guided by these principles, the Executive Order’s second section outlines six objectives for key Executive Agencies to pursue to support and ensure the Nation’s advancement in AI.
American AI Initiative Key Objectives:
- Promote cross-sectional investment initiatives in AI R&D throughout the government and with private, academic, and international stakeholders to lead to breakthrough technologies for the economy and national security;
- Enhance access to Federal data, models, and computing resources valuable to increasing the value of the country’s investment in AI R&D while also preserving the data’s lawful, safe, secure, and confidential use;
- Reduce barriers to AI’s use and innovation while preserving the Nation’s safety, economic and national security, privacy, and values;
- Minimize AI vulnerabilities while promoting trust both in AI’s operation and global use of the technology;
- Promote and provide STEM training and skills development for AI operators and innovators; and
- Implement an action plan to protect the United States AI interests, in coordination with this Executive Order.
The key implementing Executive Agencies, outlined in section three, will be coordinated through the Executive Offices’ National Science and Technology Council (NSTC) Select Committee on Artificial Intelligence. The NSTC Select Committee on Artificial Intelligence is comprised of the Federal Government’s most senior AI R&D officials and is chartered to set Federal inter-agency priorities and strategies for the technology’s development and use.
While other specific agencies are not identified by this Executive Order, whatever agencies will implement this order will be those agencies regulating, funding, developing, and/or deploying AI, as defined by the co-chairs of the NSTC Select Committee. The co-chairs of this committee are Deputy Assistant to the President for Technology Policy Michael Kratsios, National Science Foundation Director Dr. France Cordova, and Defense Advanced Research Projects Agency Director Dr. Steven Walker.
Next, in section four, implementing agencies are directed to take the objectives of this executive order into account when creating their budgets for 2020 and expending resources in 2019. In particular, the implementing agencies will:
- Coordinate with the Office of Management and Budget (OMB) and Office of Science and Technology Policy (OSTP) annually via the consortia of research and technology-focused agencies participating in the Networking and Information Technology Research and Development (NITRD) Program within the NSTC.
- Within 90 days of appropriations, the implementing agencies will report to NITRD what incoming funds will be applied to new or ongoing AI R&D initiatives; and
- Identify and collaborate with potential state, local, tribal, academic, and private AI R&D stakeholders.
Next in section five, the Executive Order calls for all entities across the Federal government to reorganize their data storage and models to be more easily accessible to non-federal researchers while preserving the security and privacy of sensitive information. The aims of this section are meant to support the application and efficiency of AI use in the government’s activities. To ensure these developments, the Executive Order establishes the following deadlines:
- Within 90 days, the OMB will publish a notice in the Federal Register to request comments from the public regarding specific data accessibility requests for AI R&D. Additionally, by this deadline, both the OMB and Select Committee on Artificial Intelligence will identify remaining data access barriers throughout the government;
- Within 120 days, the OMB, with the Select Committee, will update guidance documents on for Enterprise Data Inventories and Source Code Inventories used by government agencies;
- Within 180 days, agencies maintaining key data resources of interest to the AI R&D community with consider and implement strategies to improve the accessibility of their data while also identifying any associated resource constraints. Agency actions for this deadline will also be made in accordance with stipulations of the Cross-Agency Priority Goal: Leveraging Federal Data as a Strategic Asset, from the March 2018 President’s Management Agenda. Accordingly, discretionary high-computing resources from the Departments of Defense, Commerce, Health and Human Services, and Energy, as well as the National Aeronautics and Space Agency and National Science Foundation will be allocated to enable these strategies.
Also, within 180 days, the Select Committee on Artificial Intelligence, with the General Services Administration, will submit a report to the President with recommendations for the implementation of cloud computing technologies to improve the aims of this Executive Order.
Actions within this section will also be made in coordination with the Federal Executive Privacy Council (established by the February 9, 2016 Executive Order 13719.) Specific concerns of this collaboration include:
- Identifying and mitigating threats to individuals’ privacy and civil liberties caused by this increased access to data;
- Prioritizing safety and security concerns resulting from the association or compilation of data;
- Improving the documentation and storage of data, implementing inter-operable and machine-readable formats; and
- Considering necessary adaptations of data storage to improve its governance.
Finally, this section also stipulates that the Select Committee on Artificial Intelligence will also provide technical expertise to the American Technology Council regarding the improvement and implementation of AI-enabled services in the government and throughout society.
Section six of the Executive Order addresses the President’s policies towards the regulation of AI. Specifically, this section stipulates three deadlines for publications from various federal agencies on this topic. These stipulations include:
- That within 180 days of this Order, the OMB, with the OSTP, the Domestic Policy and National Economic councils, and other relevant agencies will publish for comments a memorandum to the Federal Register that serves as guidance for the development of regulatory and non-regulatory approaches to support the President’s guiding principles and objectives regarding AI;
- That within 180 days of the published memorandum outlined above, AI implementing agencies shall submit plans to the OMB that outline their strategies to fulfill the mandates of the memorandum;
- That within 180 days of this Order, the Department of Commerce, with the National Institute of Standards and Technology (NIST), the Select Committee on AI, and other relevant AI stakeholders, shall publish a comprehensive plan to direct Federal engagement in the creation of technical standards for the development of AI technologies consistent with OMB Circular A-119, a standing executive memo outlining requirements for the creation and monitoring of voluntary Federal consensus standards.
Under section seven of this Executive Order, the implementing AI agencies are directed to coordinate with the Select Committee on AI to prioritize funding and fellowships for AI R&D and the technology’s application in society and government. Also stipulated in this section is the stipulation that within 90 days of this order, the Select Committee on AI will advise the NSTC and National Council for the American Worker on AI-related educational and workforce development strategies for US citizens.
In the eighth and last policy section of this Order, the National Security Advisor, with the OSTP, National Security Council, and Homeland Security Council members are directed to submit a national plan to preserve the US’ in critical AI technologies with 120 days.