This month the National Security Commission on Artificial Intelligence (NSCAI) delivered its final report to the White House and US Congress. Created by the Congress in 2019, the NSCAI brought together a bipartisan group of leading US technologists, national security officials, and business leaders to provide recommendations on how the United States can lead in AI and apply the technology to key national security challenges.
The NSCAI‘s report is comprehensive, clear-eyed, and authoritative, and contains a sobering core message: “America is not prepared to defend or compete in the AI era.” Central to that stark assessment is China’s rapid catch-up in both research and deployment of AI systems. Lest there’s any doubt who the main competitor is, the mammoth 752–page report contains the words “China” or “Chinese” nearly 700 times.
But that frequent invocation of China in the report’s assessment can obscure a deeper truth found in its actionable recommendations. In the competition for global AI leadership, America’s chief opponent isn’t China, but America itself. Or more specifically, it’s the relentless partisan bickering that often grinds policymaking to a halt.
Reorganizing, Funding, and Attracting
Three core themes undergird the hundreds of concrete recommendations for AI competitiveness in the report: bureaucratic reorganization, effective industrial policy, and sustaining AI talent pipelines.
Bureaucratic reorganization is often messy, but the recommendations in this area are perhaps more achievable. Core to them is the creation of a Technology Competitiveness Council (TCC) in the mold of the National Security Council (NSC). Chaired by the Vice President, the TCC would be tasked with creating a National Technology Strategy and balancing the sometimes conflicting imperatives of commerce and security for tech competitiveness.
Industrial policy, once a taboo topic, has gained much more currency in today’s debates about technology. Here the NSCAI’s recommendations are myriad, including creating a public cloud computing infrastructure, subsidizing domestic semiconductor manufacturing, and doubling federal funding of AI R&D each year through 2026, among others.
Each of these recommendations on its own isn’t revolutionary. But taken together they signal a new conventional wisdom: to maintain AI leadership, the US government needs to play a more activist role in setting strategy and putting the money behind it.
The third central theme of the report is what it calls the “holy grail” of AI competition: talent. While the report calls for major new investments in domestic education—including passing a second National Defense Education Act—it bluntly states that the US education system “does not have the capacity nor the quality” to produce enough homegrown AI talent.
To fill the void, the report recommends ramping up highly skilled immigration: stapling green cards to all STEM PhDs earned in the United States, creating new entrepreneur visas, and doubling employment-based green cards.
The Partisanship Problem
What’s remarkable about these core recommendations is their singular focus on the American home front, with only tangential ties to US policy toward China. They are anchored in domestic policymaking to direct resources to strengthen the US technology base.
In a sign that the Biden administration will take up some of the recommendations in the report, it recently appointed NSCAI commissioner Jason Matheny to three key White House roles on technology and national security. Still, taking meaningful action in all three areas is likely to encounter major hurdles. Even as competition with a foreign adversary has been a uniting force in US politics, today’s hyper-partisan political environment can stymie smart policies.
In terms of difficulty, creating an NSC-like technology council seems the easiest and safest. Biden could sign an executive order today with little opposition. But increasing federal spending as part of industrial policy and expanding immigration are thorny.
For instance, the report’s proposed spending on bolstering semiconductors manufacturing comes to $35 billion over five years, just 1.8% of the cost of the Covid-19 relief package that was recently passed. But even those modest sums are likely to encounter opposition from deficit hawks and those ideologically opposed to anything resembling “picking winners” in free markets.
The NSCAI report attempts to preempt this, declaring “this is not a time for abstract criticism of industrial policy or fears of deficit spending to stand in the way of progress.” Early signs could be a bellwether on this front, with what appears to be bipartisan support for a bill to fund chip manufacturing.
When it comes to talent—the most important aspect of AI leadership—it requires touching the third rail in American politics at the moment: immigration. Time is not on America’s side. It simply does not have the decades needed to reinvigorate its own education system and train a fully homegrown AI workforce. Avoiding falling behind requires expanding skilled immigration from countries around the world—including China.
Unfortunately, the political divisiveness of immigration could render these arguments moot. The best hope may lie in an unspoken partisan cease fire anchored in a shared understanding that highly skilled tech immigration can be separated from the endless trench warfare of broader immigration reform. It’s a long shot, but given the importance of talent to AI competitiveness, it’s one worth taking.
Matt Sheehan is a Fellow at MacroPolo. You can find his work on tech policy, AI, and Silicon Valley here.
Get Our Stuff
Get on our mailing list to keep up with our analysis and new products.Subscribe