June 7, 2023


Skillful Technology Connoisseurs

Home » America’s global leadership in human-centered AI can’t come from industry alone

America’s global leadership in human-centered AI can’t come from industry alone

The Biden administration has followed through on a Congressional mandate to create a National AI Research Resource Task Force. With top experts from the federal government, higher education, and private organizations, the task force is dedicated to strengthening America’s foundation and spurring advances in artificial intelligence (AI). As a computer scientist, AI researcher and educator, co-director of the Stanford Institute for Human-Centered AI, and a major supporter of the bipartisan legislation that authorized this endeavor, I am honored to have accepted an offer to serve as a member of the task force. The time has never been more critical for us to come together and cement America’s leadership in AI — a technology that has the potential to drive innovation in every industry, from manufacturing and healthcare to transportation and defense.

Like all technologies human civilization has built, AI is a tool that is as good or as bad as those who make and use it. It reflects the values of our society and our time. State and non-state actors are building increasingly sophisticated technological means not only to decelerate our economy, but even to surveil citizens, censor the free speech to which Americans are accustomed, and institute other social controls. This could likely lead to a grim digital authoritarianism that can threaten people’s ability to speak, live, act, and even think freely, regardless of where one is in the world.

On the other hand, the invention of American-led technology — from lightbulbs to aircraft to personal computers and the Internet — rooted in universal human rights and democratic values can become the basis for greater freedom and prosperity. AI can manage routine functions and allow people to focus on more meaningful work. Digital tools can aid medical diagnostics and treatment, improving care for patients and helping our elderly population. AI technologies can be used to defend against ransomware and other cyberattacks to make our nation and others much more secure.

With our unique ability to scale our creative power and to use our values to shape AI for good, and an unparalleled innovative ecosystem of industry, universities, and the federal government with vast potential waiting to be unleashed, America is an integral and irreplaceable proponent of human-centered AI. To ensure technology in the future benefits society, we must lead in innovation, working with our allies to develop and use AI according to our democratic principles and informed by our longstanding respect for human rights.

With this in mind, the mission of our National AI Research Resource Task Force is defined by a clear yet difficult problem. America’s economic, scientific, and national security leadership depends upon our ability to continually generate ever more innovative breakthroughs in advanced science and technology, including artificial intelligence. Yet today, the expensive compute power and data necessary for AI research and development are available only to a small group of the most advanced technology companies. And even while these powerful technology companies successfully explore and advance AI applications, speculative basic and foundational academic research has slowed as resources have grown out of reach, and as faculty have departed for industry. With the drain of AI PhDs to commercial ventures, not only is there less research to generate long term innovation or benefit society, there are fewer people to train the next generation of academics and industry workers.

The answer is clear: to remain the global leader in AI, the United States must build a much more expansive, inclusive, and robust innovative ecosystem of industry, academia, civil society and the federal government. Key to this will be our ability to cultivate and support the vast untapped potential of American’s academic researchers who don’t currently have access to the same infrastructure of AI research and development (R&D) as large tech companies. Without federal funding of this basic and foundational AI research, which traditionally takes place at universities and then is commercialized in industry, the pipeline of AI innovation will quickly run dry.

With this imperative, our task force has been given the assignment of drafting a framework for a National Research Cloud that will balance the needs of the diverse parties involved in AI R&D and democratize AI research, education, and innovation. Not only will a National Research Cloud aid today’s researchers in leading critical, foundational research, it will also serve as an invaluable resource for teaching the next generation of AI scientists, academics, inventors, and industry leaders.

The National Research Cloud is a good and necessary start to stimulate novel research and transformative breakthroughs in artificial intelligence. Through the work of the National AI Research Resource Task Force, I hope we can bring all types of people and groups together to build a foundation for truly human-centered AI.

The future of global technology development may depend on us getting this right, and that’s not hyperbole.

Dr. Fei-Fei Li is the Sequoia Professor, Computer Science Department, and Denning Codirector, Stanford Institute for Human-Centered Artificial Intelligence, at Stanford University. She is an elected Member of the National Academy of Engineering, and the National Academy of Medicine.