When Congress unveiled a bipartisan plan to promote quantum-computing research and development this summer, many lawmakers stressed the massive surge of Chinese interest in quantum to warn that America risks falling behind in an escalating technological-arms race.
That same reasoning also applies to artificial intelligence, another emerging technology with tremendous implications for national security and economic growth. But while China and a host of other nations announce sweeping plans to seize global leadership on AI, Congress continues to deploy only piecemeal initiatives that focus on national security over civilian applications.
An aide to the House Science Committee, which is spearheading the push for a federal quantum initiative in the lower chamber, said Capitol Hill’s sudden fixation on quantum research is unlikely to lead to the development of a similar framework on AI.
“The horse has already left the barn, in some sense, on AI,” the aide told National Journal. “The commercial applications are so broad, the private sector is already investing heavily in it. And I think that any attempts by government to try and intervene could be constraining on the development of this technology.”
Frederick Hill, a spokesman for the Senate Commerce Committee chaired by Sen. John Thune, noted that "there are ongoing conversations related to AI but no timeline or immediate plans for moving legislation."
But while the American AI industry is now flush with cash and driven by real-world opportunities, experts say it’s unlikely to retain its global advantage absent the creation of a comprehensive partnership between the federal government and private actors. They believe a national framework is needed to reform immigration laws to attract top-tier programmers, invest in STEM education at home, fund new applications that increase AI safety, and build a set of guidelines for ethical AI use—all vital components of a thriving AI economy, and ones that are likely to falter without federal leadership.
“If we sit back and let things take their course, China will take the lead in artificial intelligence in the next 15 years,” said Paul Scharre, the director of the technology and national security program at the Center for a New American Security.
And unlike quantum, where dramatic breakthroughs remain several years away at best, experts say AI’s immediate impact on both the civil and national security sectors makes it an issue that demands even greater urgency.
“I think the need for Congress to develop an AI strategy is even more pressing, just because of the scale and the immediacy of AI compared to quantum computing,” said Joshua New, a policy analyst at the Center for Data Innovation.
Experts in emerging technologies are the first to admit serious differences between the quantum-computing and artificial-intelligence industries. Real-world applications for the spooky science of quantum remain nebulous, and the tremendous cost of creating quantum-computing hardware means the machines are currently of dubious economic value to the private sector.
Rapid advances over the past 10 years in machine learning, autonomous vehicles, and voice-assistance technology have meanwhile transformed the AI industry into a money-making powerhouse for U.S. tech companies large and small.
“You can start an AI company at a seed stage for around a million dollars,” said Evanna Hu, the chief executive of machine-learning company Omelas and a fellow at New America. “Quantum right now, if you don’t raise more than $25 million for a seed-stage company, then there’s no way that you can actually stand things up.”
As an expert on both AI and quantum computing, Hu says she’s long advocated for increased government support for both emerging technologies. But in her interactions with public officials in recent months, she’s noticed a slump in the interest surrounding AI.
“I question sometimes if AI is not really—if the fad is kind of fading away, and quantum is in,” Hu said.
It’s not that the federal government never tried to build a unified AI strategy. Hu said the Defense Advanced Research Projects Agency worked relentlessly throughout the 1990s to attract corporate partnerships on AI research. “The private sector just never really picked it up,” Hu said, adding that companies were at the time more focused on miniaturizing personal computers.
Another shot came at the tail end of the Obama administration, when the White House released a report on AI that experts believe clearly teed up the development of a comprehensive strategy with buy-in from Congress. But that momentum died with the election of President Trump.
“It was really Obama directly that was pushing a lot of this stuff ahead,” said Tim Hwang, the director of the Harvard-MIT Ethics and Governance of Artificial Intelligence Fund. “I don’t think we see a similar interest here.”
The lack of interest on Capitol Hill is compounded by big tech’s own skepticism of Washington’s priorities. While companies like IBM and Intel are eager to begin working with the federal government on quantum, Silicon Valley is increasingly wary of partnering with the U.S. government on AI. Earlier this summer, Google decided against renewing a Pentagon contract to develop AI that can identify drone targets from satellite images. And in July, more than 2,400 AI researchers around the world—including many top scientists in the United States—signed a pledge to never use their expertise to develop lethal autonomous weapons systems.
Military and defense applications are the main areas where Congress has sought to widen its AI footprint—lawmakers authorized millions of dollars in AI research funding in the just-passed National Defense Authorization Act. That focus, many experts argue, is counterproductive when it comes to achieving support from top-tier AI researchers.
“I think this is where there’s a real need for a national strategy,” said Hwang, who stressed the importance of an ethical AI framework crafted by the federal government and focused on the technology’s peacetime applications.
“It actually ends up being really key if we think about it in terms of competitiveness,” said Hwang, “not just in terms of how the technology is deployed, but also like, are researchers willing to cooperate with the government on these issues?”
The House Science Committee aide echoed that assertion. While still resistant to the idea of an overarching national strategy, the aide noted that the committee voted for a reauthorization bill this summer that gives the National Institute of Standards and Technology $40 million to develop ethical AI guidelines. The aide also stressed that if a national AI strategy ever does materialize, it’s important that it be a “civilian-led effort.”
Given the glut of private-sector funding, America’s AI supremacy isn’t in danger of imminent collapse. But without a national strategy to retain and attract capital and programming talent, experts say the U.S. advantage remains in jeopardy. While U.S.-based companies absorbed 66 percent of AI investments in 2016 to China’s 17 percent, by 2017 Chinese companies had jumped to 48 percent of total global AI investments. American firms attracted just 38 percent.
“[The U.S.] lead is slipping,” said New. “And I think without action now—this shouldn’t be something the U.S. responds to after we’ve already lost our competitive edge. It has to be something that the government is proactive about maintaining.”