Many of the most fundamental breakthroughs powering the U.S. economy today—the laser, gene sequencing, the Internet and GPS, to name a few—were made possible by federal funding for basic research carried out by universities. The productive interplay of the federal government, research universities, and private enterprise has given rise to an American innovation engine that is the envy of the world.
Artificial intelligence, one of the most consequential technologies of the 21st century, is a direct outgrowth of federally-funded university research, further advanced by exceptional R&D in the private sector. Here, our innovation engine has run true to form. But today, the research prowess that’s powered decades of growth and prosperity is at risk.There are two reasons: public researchers’ lack of access to compute power and the scarcity of meaningful datasets, the two prerequisites for advanced AI research.
The United States has been and remains a leader in AI research. But declining government investment in basic and foundational research, combined with lack of access to computational resources and large datasets, threatens America’s position on the global stage. This is a disservice to the public, to our economy and to national security. Today’s AI requires vast amounts of compute, data and expertise to train and deploy the massive machine learning models powering the most advanced research. There is a wide gulf between the few companies that can afford these resources and everyone else. To put this in perspective, Google used nearly $1.5 million in compute cycles to train the Meena chatbot it announced earlier this year. Such costs for a single research project are out of reach for most corporations, let alone for academic researchers. At present, only a handful of companies can afford the substantial computational resources required to develop and train the machine learning models underlying today’s AI. What’s more, the large data troves required to train these algorithms are for the most part controlled by either industry or government. Academic researchers struggle to gain access to both, which results in a hobbling of this crucial partner in the American research enterprise.
It is for this reason that we are calling for the creation of a U.S. Government-led task force from academia, government, and industry to establish a National Research Cloud. Support from Congress and the President could have a meaningful impact on American innovation through the creation of such a task force. Indeed, we believe that this could be one of the most strategic research investments the federal government has ever made.
The National Research Cloud will be a close partnership between academia, government, and industry. It will provide academic researchers with affordable access to high-end computational resources, to large-scale government-held datasets in a secure cloud environment, and to the necessary expertise to benefit from this resource. In doing so, it will unleash a pent-up supply of public-interest AI research, while accelerating the training of a new generation of AI engineers at universities throughout the country.
Stanford HAI launched this initiative last year, organizing the presidents and provosts of 22 of the top universities across the country to sign on to a joint letter to the President and Congress in support of this effort. As the leadership of our nation’s universities must now focus on their campus’ response to the pandemic, we decided to hold off on issuing the joint letter. But in spite of the challenging environment we currently face, it is important to move this initiative forward. In fact, it is times like these when research is most needed to address current and future challenges.
This joint initiative will equalize access to these critical resources across the country, elevating the ability of all colleges and universities to provide the research and teaching needed to maintain our competitiveness in AI. America must develop a deep bench of talent to conduct research that will lead to the next generation of technological breakthroughs and to build applications of AI that benefit all of humanity. Perhaps most important, we must ensure that these benefits remain open and are shared by all.
Imagine for example what public interest researchers across the country could do to improve the health of Americans while reducing healthcare costs if the U.S. government unlocked privacy-protected access to massive datasets such as Medicare, TRICARE, and the Veterans Administration, matched with an investment in computational resources. Or think of how government-held weather data could be used to forecast crop yields, better predict dangerous storms, and prepare affected communities further in advance. The list goes on and on.
We have seen only the beginning of what artificial intelligence can provide to enhance the quality of human life, to address the challenges we face as a nation and world, and to advance our knowledge of the universe around us. The best way to ensure that the technology progresses and is put to human-centric uses is to democratize access to the essential tools of the trade, empowering public-minded academic and educational institutions to participate in this endeavor. A National Research Cloud will do exactly that.