A group of more than 20 organizations including tech giants like AWS, Google, IBM, and Nvidia joined schools like Stanford University and The Ohio State University today in backing the idea of a national AI research cloud. Nonprofit groups like Mozilla and the Allen Institute for AI also support the idea. The cloud would help researchers across the United States gain access to compute power and data sets freely available to companies like Google, but not researchers in academia. Compute resources available to academics could grow even more scarce in the near future as COVID-19 fallout constricts university budgets.
The National AI Research Resource Task Force Act was first introduced earlier this month by founding cochairs of the Senate AI Caucus U.S. Senators Rob Portman (R-OH) and Martin Heinrich (D-NM) together with a bipartisan group in the House of Representatives. If passed, the bill will bring together experts from government, industry, and academia to devise a plan for the creation of a national AI research cloud.
The National Security Commission on Artificial Intelligence (NSCAI) chairman and former Google CEO Eric Schmidt also supports the plan. In reports written by tech executives and delivered to Congress in the past year, the NSCAI has recommended more cooperation between academia, industry, and government as part of a broader strategy to keep the United States edge in tech compared to other nations.
The idea of a national AI research cloud was first proposed last year by the Stanford Institute for Human-Centered Artificial Intelligence (HAI) codirectors Dr. Fei-Fei Li and John Etchemendy, calling its creation essential to U.S. competitiveness and the nation’s status as a leader in AI. In a March blog post, Li and Etchemendy called the creation of such a cloud potentially “one of the most strategic research investments the federal government has ever made.”
Last year, leaders at Stanford joined more than 20 other universities in sending a joint letter to President Trump and Congress last year backing a national AI research cloud. Previous bills like Senator Heinrich (D-NM) also recommended the creation of AI centers and a national AI coordination office as part of a comprehensive U.S. AI strategy. Increased data sharing and ideas like a national center of excellence were also urged by researchers last year when the Computing Community Consortium laid out its 20-year AI research road map.
Li talked about AI, China, health care, and other topics today in a conversation with former Secretary of State and soon-to-be Hoover Institution director Condoleezza Rice.
After stating that a U.S. lead in tech is important to national security, Rice asked Li about how the U.S. can lead in AI if China has more data and fewer privacy concerns. In response, Li said AI like speech or facial recognition may be data heavy, but other forms of AI that require less data may be supply fruitful ground for U.S. progress.
“Data is a first-class citizen of today’s AI research. We should admit that, but it’s not the only thing that defines AI,” Li said. “Rare disease understanding, genetic study of rare disease, drug discovery, treatment management — they are by definition not necessarily data heavy, and AI can play a huge role. Human-centered design, I think about elder care and that kind of nuanced technological help. That’s not necessarily data heavy as well, so I think we need to be very thoughtful about how to use data.”
The future of work, ethics, and AI bias were also major topics of discussion. Li urged the development of AI that brings together interdisciplinary teams, gathers insights from people impacted by AI, and is made by more than computer science school graduates.
“America’s strength is our people and the more people who participate in this technology, to guide and develop it, the stronger we are,” she said.
Li also stressed the need to stay ahead of the ethical implications of AI and suggested computer scientists throw away the notion of independent machine values, asserting that “Machine values are human values.”
In a separate policy proposal made by Stanford HAI last year, Li and Etchemendy urged the federal government to grow its national AI investments to $12 billion a year for the next decade.