On March 19, 2019, the US federal government launched AI.gov to make it easier to access all of the governmental AI initiatives currently underway. The site is the best single resource from which to gain a better understanding of US AI strategy. As of February 2020, there is also extensive information, data, and graphics about AI policy in the US available at the OECD AI Policy Observatory.
The following description of AI policy in the US provides a snapshot summary of some previous and ongoing initiatives throughout the country. It is regularly updated, but it is not intended to be comprehensive.
American AI Initiative
US President Donald Trump issued an Executive Order launching the American AI Initiative on February 11, 2019. The Executive Order explained that the Federal Government plays an important role not only in facilitating AI R&D, but also in promoting trust, training people for a changing workforce, and protecting national interests, security, and values. And while the Executive Order emphasizes American leadership in AI, it is stressed that this requires enhancing collaboration with foreign partners and allies.
The American AI Initiative is guided by five principles, which include (in summarized form), the following: 1. Driving technological breakthroughs, 2. Driving the development of appropriate technical standards, 3. Training workers with the skills to develop and apply AI technologies, 4. Protecting American values including civil liberties and privacy and fostering public trust and confidence in AI technologies, 5. Protecting US technological advantage in AI, while promoting an international environment that supports innovation.
The Executive Order calls on the National Science and Technology Council (NSTC) Select Committee on Artificial Intelligence to coordinate the American AI Initiative. All executive departments and agencies that are developing or deploying AI, providing educational grants, or regulating or guiding AI are required to adhere to six strategic objectives. These objectives (in summarized form), include: promoting sustained investment in AI R&D; enhancing access to Federal data, models, and computing resources; reducing barriers to the use of AI technologies; ensuring that technical standards minimize vulnerability to attacks from malicious actors; training American AI researchers; and implementing an action plan to protect US economic and national security interests.
The heads of each agency are requested to increase focus on and budget for AI into the next Fiscal Year, and every year thereafter. They are also required to communicate their plans to achieve this prioritization to the OMB Director and the OSTP Director each fiscal year through the Networking and Information Technology Research and Development (NITRD) Program. Each agency will also be tasked with developing regulatory and non-regulatory approaches to guide AI development and use, following the release of a memorandum expected 180 days following the Executive Order. Moreover, the Director of the National Institute of Standards and Technology (NIST), is tasked with creating a plan for Federal engagement in the development of technical standards to support reliable, robust, and trustworthy AI systems. NIST put out a Request for Information on the topic in May 2019, and the comments that were received can be viewed here.
The day after the Executive Order was released, the US Department of Defense followed up with the release of an unclassified summary of its own Artificial Intelligence Strategy. The U.S. Air Force released an Annex to this strategy to share its own 2019 Artificial Intelligence Strategy in September 2019.
International engagement is a key pillar of the American AI Initiative.
In May 2019, the United States joined dozens of other countries in adopting the OECD AI Recommendation, the first intergovernmental standard for AI, which includes five complementary values-based principles and five recommendations to governments. The following month the US also joined the G20 countries in supporting the G20 AI Principles, which are drawn from the OECD Recommendation. In September 2019, the US Chamber of Commerce released Principles on Artificial Intelligence, which also endorse the OECD Recommendation and include a call for US businesses to abide by international standards.
Federal Funding Landscape
AI R&D is a top priority for the US and has enjoyed widely bipartisan support. In June 2019, the Select Committee on Artificial Intelligence and the National Science and Technology Council released an updated version of the National AI R&D Strategic Plan, which includes eight strategies to guide the portfolio for Federal AI R&D investments. In September 2019, NITRD released a Supplement to the President’s FY2020 Budget, which provided the first agency-by-agency budget breakdown for non-defense AI R&D, showing a total budget of nearly $1 billion for the year.
In a July 31, 2018 memo from the Executive Office of the President, leadership in AI (along with “quantum information sciences and strategic computing”) is named the second highest R&D priority after the security of the American people for the fiscal year 2020. And on September 7, 2018, the U.S. Department of Defense announced it will invest up to $2 billion over the next five years towards the advancement of AI. This is in addition to existing government spending on AI R&D, which totaled more than $2 billion in 2017 alone, just from unclassified programs and not including Pentagon and intelligence budgets. Existing funding has already propelled more than 20 active programs under the Defense Advanced Research Projects Agency (DARPA) exploring different aspects and uses of AI, and dozens of new projects have now been promised.
This funding followed the announcement of a National Security Commission on Artificial Intelligence, which was made official with U.S. President Trump’s signing of the 2019 National Defense Authorization Act (NDAA) in August 2018. The Commission includes 15 members selected by different government officials and is tasked with the assessment of the national security implications of AI including the ethical considerations of AI in defense (the Commission released its Interim Report November 2019.) The Department of Defense (DoD) additionally established a Joint AI Center (JAIC) in July 2018 to explore the agency’s use of AI. The JAIC works on AI “National Mission Initiatives;” improving collaboration with the private sector, academia, and military allies; attracting AI talent and establishing an ethical framework for AI in defense; and aiding in the National Defense Strategy. The JAIC is tasked with carrying out the DoD’s AI Strategy. The Defense Innovation Board released AI Principles to provide ethical recommendations to the DoD in November 2019.
President Trump was also the first U.S. president to specifically name artificial intelligence as an Administration R&D priority in his 2019 Budget Request to Congress. AI was also featured for the first time in the 2018 National Security Strategy in relation to its role in helping the U.S. lead in technological innovation as well as its role in information statecraft, weaponization, and surveillance. It also shows up for the first time in the 2018 National Defense Strategy where it is described as one of the technologies that will change the character of war and give increasingly sophisticated capabilities to our adversaries, including non-state actors.
Other governmental initiatives are noteworthy as well. In July 2017, the Department of Homeland Security put out a report titled, “Artificial Intelligence Risk to Critical Infrastructure,” which analyzes narratives about AI to better understand the perception of benefits and threats from AI adoption.
White House Summit on AI and the AI Select Committee
In May 2018, President Trump and the White House held a Summit on Artificial Intelligence for American Industry that included key technology companies. The White House also released a Fact Sheet titled, “Artificial Intelligence for the American People,” which highlighted the Trump Administration’s priorities for AI. Trump declared, “To the greatest degree possible, we will allow scientists and technologists to freely develop their next great inventions right here in the United States.” The priorities discussed were: funding AI research, removing regulatory barriers to the deployment of AI-powered technologies, training the future American workforce, achieving strategic military advantage, leveraging AI for government services, and working with allies to promote AI R&D.
The White House announced plans to help provide U.S. companies with new data sources, and to establish a Select Committee on Artificial Intelligence to help government agencies think about and use the technology, as well as consider partnerships with industry and academia. Notes from the Committee’s inaugural meeting June 27, 2018 can be found here. At the meeting, the Committee and the Office of Science and Technology Policy (OSTP) approved the formation of a new Networking and Information Technology Research and Development (NITRD) Interagency Working Group (IWG) for AI R&D. The IWG will serve as the coordinating body for Federal AI R&D investments and will report to both the NITRD Subcommittee and the Machine Learning and AI (MLAI) Subcommittee.
In September 2019, the White House held another AI Summit to address the use of AI in government. Key takeaways of the Summit included sharing learning between government, industry, and academia; developing a Center of Excellence model for government agencies to share AI expertise and best practices; and discussion of ways to build AI skills within the government workforce.
AI in Congress
The bipartisan Artificial Intelligence (AI) Caucus for the 115th Congress launched in May 2017, originally co-chaired by Congressman John K. Delaney and Congressman Pete Olson. According to the press release from Delaney’s office at the time, “The goal of the caucus is to inform policymakers of the technological, economic and social impacts of advances in AI and to ensure that rapid innovation in AI and related fields benefits Americans as fully as possible.” On November 7, 2017, the AI Caucus co-hosted a luncheon briefing to discuss ethics and privacy issues related to artificial intelligence with the Software & Information Industry Association and IEEE-USA, called, “Machines That Learn: Can They Also Be Taught Human Values?”
Additionally, many bills have been introduced in Congress that mention or focus on artificial intelligence. There are numerous bills that relate to autonomous driving, including The SELF DRIVE Act, which passed the House in September 2017. The bill requires the Department of Transportation (DOT) to undertake research on the best way to inform consumers about the capabilities and limitations of highly automated vehicles.
Other bills relate to the economic impacts of AI. For example, the Innovation Corps Act of 2017, introduced in the House on March 16, 2017, would help retrain workers. It states, “The acceleration of artificial intelligence is enabling the automation of jobs that have traditionally required human labor.” The AI JOBS Act of 2018, introduced in the House on January 18, 2018, states, “It is the sense of Congress that technology can improve the lives of individuals, but can also disrupt jobs, and for this reason, innovation should be encouraged while training and retraining American workers for our 21st century economy.” The Act would require the Secretary of Labor to prepare a report on artificial intelligence and its impact on the workforce.
Bills that address the ethical implications of AI have also been introduced. For example, House Resolution 153 (HRES 153) was introduced February 2019 to support the development of guidelines for ethical development of artificial intelligence in consultation with diverse stakeholders, and consonant with the aims of: (1) Engagement among industry, government, academia, and civil society. (2) Transparency and explainability of AI systems, processes, and implications. (3) Helping to empower women and underrepresented or marginalized populations. (4) Information privacy and the protection of one’s personal data. (5) Career opportunity to find meaningful work and maintain a livelihood. (6) Accountability and oversight for all automated decisionmaking. (7) Lifelong learning in STEM, social sciences, and humanities. (8) Access and fairness regarding technological services and benefits. (9) Interdisciplinary research about AI that is safe and beneficial. (10) Safety, security, and control of AI systems now and in the future.
The FUTURE of AI Act 2017 was introduced in both the House and Senate on December 12, 2017. It would require the Secretary of Commerce to establish a Federal Advisory Committee on the Development and Implementation of Artificial Intelligence. This committee would consider US competitiveness, workforce and technological displacement, education, ethics, data sharing, international cooperation, accountability and legal rights, cultural and social norms, impact on rural communities, and government efficiency. The Algorithmic Accountability Act was introduced April 2019 and would direct the Federal Trade Commission to require entities that use, store, or share personal information to conduct automated decision system impact assessments and data protection impact assessments.
Throughout the beginning of 2018, The Subcommittee on Information Technology led a three-part series on artificial intelligence: In February the subcommittee held the first hearing, which was considered a chance to build a foundation of knowledge about what AI is and how it is being used. The second hearing was held in March and featured government witnesses who shared how their agencies are engaging with AI through research, procurement, and applications. The third hearing was held in April and considered the challenges posed by AI and the range of potential policy responses, including the appropriate role for government. Additionally, the U.S. House of Representatives Subcommittees on Research and Technology and Energy held a hearing titled “Artificial Intelligence – With Great Power Comes Great Responsibility,” on June 26, 2018 in which they considered “artificial general intelligence”.
There have been numerous other Congressional hearings related to AI. For example, the House Intelligence Committee held a hearing on “The National Security Challenge of Artificial Intelligence, Manipulated Media, and “Deepfakes” in June 2019. Later that same month, the U.S. House Committee on Science, Space, & Technology held a hearing on the societal and ethical implications of AI.
AI Under the Obama Administration
Steps towards AI policy in the U.S. were first seen under the Obama administration. In 2016, the former President and the White House OSTP launched a series of workshops and a Subcommittee on Machine Learning and Artificial Intelligence to monitor technological advances and help coordinate federal activity in AI. These activities led to the creation of three globally influential reports: “Preparing for the Future of Artificial Intelligence,” “The National Artificial Intelligence Research and Development Strategic Plan,” and “Artificial Intelligence, Automation, and the Economy.”
State and Local AI Policy
There are many AI-related bills being introduced at the state and local levels. For example, in August 2018 the California State Senate passed a resolution in support of the Asilomar AI Principles – a set of 23 guidelines for the safe and beneficial development and use of AI. California also passed a bot disclosure law that makes it unlawful to use a bot without disclosing that it is not a human if the intent of the bot is to incentivize a commercial transaction or to influence a vote in an election. The Consumer Privacy Act, passed in June 2018, also requires informing people about how their personal information is being used and allowing people to opt-out of having their data sold to a third party. Numerous other AI-related bills were introduced in California in the first months of 2019.
Initiatives in other states include an AI Task Force developed in Vermont in May 2018 to make recommendations about government use of AI and state regulation of the field, and a Future of Work Task Force in Washington created in March 2018 to navigate automation and shifting skills requirements.
City level policies include a ban on facial recognition technology use by city agencies passed in San Francisco May 2019. Additionally, the New York City Council passed an algorithmic accountability bill in 2017 that established the New York Algorithm Monitoring Task Force. The group studies how city agencies use algorithms to make decisions that affect New Yorkers’ lives.
Additional Links and Resources
- “Report on Algorithmic Risk Assessment Tools in the U.S. Criminal Justice System,” Partnership on AI, April 2019.
- “JAIC: Pentagon debuts artificial intelligence hub,” Jade Leung, Sophie-Charlotte Fischer, Bulletin of the Atomic Scientists, August 8, 2018.
- “A National Machine Intelligence Strategy for the United States,” William Carter, Centre for Strategic & International Studies, 2018.