Significant contributions were provided by CRA Director of Government Affairs Peter Harsha and CCC Director Ann Schwartz Drobnis.  

Yesterday at a conference of the National Security Commission on AI (NSCAI) in DC, a bipartisan collection of Congressional and agency leaders spoke of the urgent need for the United States to retain its leadership role in Artificial Intelligence in the face of dramatically increased competition from U.S. adversaries. Senate Minority Leader Chuck Schumer (D-NY), noting Russian President Vladimir Putin’s declaration that the nation that leads AI will rule the world, affirmed the priority he believes AI research and education ought to enjoy by teasing a legislative proposal that would create a new subsidiary agency to the National Science Foundation and DARPA called the National Science and Technology Foundation (NSTF). The new agency would be endowed with $100 billion worth of federal funding over the next five years to focus on research and education in AI, Quantum Computing and other emerging technology areas. The bill, which Schumer says has broad support including among those “close to” the President and Senate Majority Leader Mitch McConnell, would radically increase the amount of federal funding for fundamental AI research in the U.S., now estimated at about $1 billion per year. 

Schumer shared few other details about the bill, which is in draft form and has not yet been introduced, but urged attendees at the conference to reach out to the White House and the Senate Majority and affirm their support. 

Tuesday’s conference served to introduce NSCAI’s much anticipated interim report to Congress (released on Monday). The NSCAI was chartered by Congress in the FY19 National Defense Authorization Act and tasked with helping guide Department of Defense efforts in Artificial Intelligence. The report is an assessment of the current state of the U.S. efforts in artificial intelligence, particularly as it relates to national security, and in it the NSCAI commissioners expressed confidence that the U.S. is “up to the challenge” of continued AI innovation leadership. The report represents the conclusion of the “analysis phase” of the commission’s work, which began earlier this year. The commission plans to spend much of the next year coming up with specific policy recommendations in a final report based on its analysis. In this initial analysis, the commission members find that the US is not translating broad national AI strengths and strategy statements into specific national security advantages: 

“Key departments and agencies have not yet fully embraced high-level strategy pronouncements and therefore critical national security missions have not incorporated AI. Given the general support for AI initiatives within government, one purpose of this interim assessment is to understand and explain the multiple reasons why that is so.”

NSCAI commissioners and other federal policymakers amplified those findings at the conference yesterday, including remarks from Energy Secretary Rick Perry, Secretary of Defense Mark Esper, former Secretary of State Henry Kissenger, Senators Chuck Schumer (D-NY) and Joni Ernst (R-IA), and Representatives Elise Stefanik (R-NY) and Kevin McCarthy (R-CA). All reiterated the urgent need for the U.S. to be the world leader in AI and many drew parallels between the current needs and the forces that drove the Manhattan Project 75 years ago. Energy Secretary Rick Perry noted the U.S. is in a race to develop AI against “freedom’s foes” and said AI is a part of the “new security triad” consisting of “exascale computing, AI, and quantum computing” that will ensure national security in much the same way the nuclear triad (strategic bombers, land-based ICBMs, and submarine launched missiles) has done for 70 years. 

Many also pointed out the importance of U.S. leadership in AI in helping ensure that U.S. ethics and values inform AI standards and not those of our adversaries. The threat of Chinese dominance in the space was highlighted frequently, but a number of speakers — including NSCAI commissioners Andrew Moore (Google), Eric Schmidt (Alphabet) and Eric Horvitz (Microsoft) — expressed caution about letting fear of Chinese competition overly impact the many areas where foreign collaborations in fundamental research are remarkably productive and mutually beneficial. 

The commission’s report explores five particular “lines of effort” relating to the nation’s AI activities that the Federal Government ought to pursue for “successful maximization of AI for national security purposes.” While the commission’s focus is AI’s implications for national security, strengthening AI efforts along these lines will likely also have a significant impact on a wide array of AI application areas beyond national security, including health care, transportation, scientific discovery and many others.

1. Invest in AI Research and Development
   • Federal R&D funding for AI has not kept pace with the revolutionary potential it holds or with aggressive investments by competitors. Investments that are multiple times greater than current levels are needed.
   • Untapped opportunities exist to build a nationwide AI R&D infrastructure and encourage regional innovation “clusters.” Such AI districts for defense would benefit both national security and economic competitiveness.
   • Bureaucratic and resource constraints are hindering government-affiliated labs and research centers from reaching their full potential in AI R&D.
2. Apply AI to National Security Missions
   • Implementation of the government’s security strategies for AI is threatened by bureaucratic impediments and inertia. Defense and intelligence agencies must urgently accelerate their efforts.
   • Implementation of the government’s security strategies for AI is threatened by bureaucratic impediments and inertia. Defense and intelligence agencies must urgently accelerate their efforts.
   • AI adoption and deployment requires a different approach to acquisition.
   • Rapidly fielding AI is an operational necessity. To get there requires investment in resilient, robust, reliable, and secure AI systems.
   • AI is only as good as the infrastructure behind it. Within DoD in particular this infrastructure is severely underdeveloped.
3. Train and Recruit AI Talent
   • National security agencies need to rethink the requirements for an AI-ready workforce. That includes extending familiarity with a range of relevant AI technologies throughout organizations, infusing training on the ethical and responsible development and fielding of AI at every level, and spreading the use of modern software tools.
   • DoD and the IC are failing to capitalize on existing technical talent because they do not have effective ways to identify AI-relevant skills already present in their workforce. They should systematically measure and incentivize the development of those skills.
   • The U.S. government is not fully utilizing civilian hiring authorities to recruit AI talent. Agencies need to make better use of pipelines for people with STEM training.
   • Expanding AI-focused fellowships and exchange opportunities can give officials and service members access to cutting-edge technology, and bring talent from our top AI companies into federal service.
   • Two realities about the American AI talent pool have become clear to us:
     1. Colleges and universities cannot meet the demand for undergraduate student interest in AI and computer science generally.
     2. The American AI talent pool depends heavily on international students and workers. Our global competitiveness hinges on our ability to attract and retain top minds from around the world.
4. Protect and Build Upon U.S. Technology Advantages
   • The U.S. government should continue to use export controls––including multilateral controls––to protect specific U.S. and allied AI hardware advantages, in particular those in semiconductor manufacturing equipment.
   • Traditional item-based export controls and narrowly-scoped foreign investment reviews are by themselves insufficient to sustain U.S. competitiveness in AI.
   • The United States must continue leading in AI-related hardware, and ensure the government has trusted access to the latest technologies.
   • Law enforcement and academic leaders can and should find common ground on preserving an open research system while reducing security risks from foreign government-directed activity on American campuses.
5. Marshal Global AI Cooperation
   • The United States must enhance its competitiveness in AI by establishing a network of partners dedicated to AI data sharing, R&D coordination, capacity building, and talent exchanges.
   • AI presents significant challenges for military interoperability. If the United States and its allies do not coordinate early and often on AI-enabled capabilities, the effectiveness of our military coalitions will suffer.
   • The United States should lead in establishing a positive agenda for cooperation with all nations on AI advances that promise to benefit humanity.

For more information, please see the interim report here. The NSCAI Conference was livestreamed and archived, and the entire program is viewable here. (Sen. Schumer’s comments begin at around 4:20:54.)