Congress asked us to serve on a bipartisan commission of tech leaders, scientists, and national security professionals to explore the relationship between artificial intelligence (AI) and national security. Our work is not complete, but our initial assessment is worth sharing now: in the next decade, the United States is in danger of losing its global leadership in AI and its innovation edge. That edge is a foundation of our economic prosperity, military power and ultimately the freedoms we enjoy.
As we consider the leadership stakes, we are struck by AI’s potential to propel us towards many imaginable futures. Some hold great promise; others are concerning. If past technological revolutions are a guide, the future will include elements of both.
Some of us have dedicated our professional lives to advancing AI for the benefit of humanity. AI technologies have been harnessed for good in sectors ranging from health care to education to transportation. Today’s progress only scratches the surface of AI’s potential. Computing power, large data sets, and new methods have led us to an inflection point where AI and its sub-disciplines (including machine vision, machine learning, natural language understanding, and robotics) will transform the world.
At the same time, we are concerned that AI is being used in ways that are antithetical to American values. In China, AI is used as a tool for centralizing power at the expense of individual rights. The Chinese government is amassing the personal data of its people, using facial recognition software to stifle dissent and repress minorities, and exporting its surveillance technology abroad.
AI also presents new threats to our security. China and Russia are investing in AI capabilities to achieve military advantages. Russia will use AI to strengthen its hybrid warfare toolkit to interfere in other states. It has used armed robotic vehicles with autonomous features in Syria. AI will accelerate and expand cyber threats to our critical infrastructure. AI will amplify disinformation campaigns and create new problems like computer-generated deepfakes designed to undermine truth.
The competitive international environment must inform our analysis. The United States is in a strategic and technological competition with China. China intends to lead the world in AI and related technologies on the way to becoming the world’s preeminent power. While the United States has long enjoyed a technical lead in AI, by many measures the U.S. lead is dwindling in research and applications. China has been aggressively investing in education and research, and leveraging advances being made openly within the worldwide AI research community. Beyond open research, China’s actions to extract knowledge and knowhow through espionage and other means is extensively documented. Its commercial sector is formidable and directly connected to its military development and strategic ambition.
Balancing the need to compete with the benefits of cooperation will not be easy. Chinese nationals make tremendous contributions at American universities and companies. Deep commercial ties exist between U.S. and Chinese tech industries. We must find a responsible equilibrium that advances our interests and our values.
What is at stake, then, in the question of leadership is the type of future we and our friends in the world want to pursue and defend.
After more than 150 discussions with AI experts ranging from skeptics to evangelists, we have identified several propositions to help the United States and its allies fashion an AI future more consistent with our shared interests and ideals.
We must lead in AI innovation. That requires serious long-term federal investments in basic research and niche security applications that the commercial sector may not pursue. AI companies are eager to do good, but the government has the ultimate responsibility for our nation’s future.
Applying AI to national security missions is an imperative. AI will change the way intelligence agencies make sense of the world, how we defend our country, and how our military fights. AI will sort and integrate videos and other information streams and will help discern patterns and pick out threats with greater accuracy than a human. AI will enable our commanders to act on the battlefield faster, with greater situational awareness, and more precision. Doing so can put fewer service members in danger and protect civilians.
To realize AI’s potential, the government and the tech sector must forge a common commitment to protecting our values and security. The tech industry generates many of the major breakthroughs, and companies are on the frontlines of defending against cyber threats and malicious uses of AI. Industry must help government discern trends and act against foreign threats. The government, for its part, should articulate clear standards and policies for responsible use, rebuild trust through greater transparency, and offer a vision of shared purpose.
We must keep the rule of law and ethics at the forefront of AI design, development and use. AI systems must be safe, robust, reliable and free of unwanted bias. To be operationally effective and ethically acceptable, systems must be trustworthy and humans must remain responsible for the outcomes of their use. We believe these goals can be achieved without paralyzing the development of AI for national security purposes or compromising our ethical standards.
Finally, in the age of AI, people will matter more than ever. AI requires critical science and engineering skills and the talent shortage is real. We must cultivate homegrown talent by making long-term investments in STEM education. We are also in a global talent competition that we must win to recruit and retain the best minds in the world. Our innovation culture, tradition of open and free inquiry, and top universities and tech companies give us advantages.
Some argue that only a “Sputnik” moment will wake the American people and government to act with purpose, just as the 1957 Soviet launch of a satellite catalyzed new educational and technological investments. We disagree. We have been struck by the broad, bipartisan consensus in America to “get AI right” now. We are in a rare moment when challenge, urgency, and consensus may just align to generate the energy we need to extend our AI leadership and build a better future.
Eric Schmidt is chairman of the National Security Commission on Artificial Intelligence (NSCAI), an independent federal commission created in 2018 to recommend steps to advance AI’s development and its national security and defense implications. He previously was CEO of Google from 2001 to 2011, executive chairman of Google from 2011 to 2015, and executive chairman of Alphabet Inc. from 2015 to 2017.
Bob Work is the NSCAI’s vice chairman and was deputy secretary of Defense from 2014 to 2017. He was the under secretary of the Navy, 2009-2013, and is the former CEO of the Center for a New American Security.