In a bold step to prevent China from getting access to vital new US technologies, last November the Trump administration announced its first move towards imposing new controls on exports of “emerging” technologies deemed “essential to national security” in categories ranging from artificial intelligence to robotics and quantum computing. The administration has called for comments from industry and academic experts on its new policy.
Though the administration’s goal is understandable as China surges ahead to try to catch up to the United States in these areas, the proposals raise concerns about the potential to weaken the global, open model of technology innovation. Overly broad or restrictive controls to “protect” US technology could strangle its development without making the United States more secure, especially if the technologies affected are not clearly linked to national security or are readily available from other countries. The US proposals could also harmfully ban Chinese nationals from working on sensitive technologies in the United States. On the other hand, excessively narrow controls could allow US-developed technologies that will form the backbone of military supply chains, surveillance, and intelligence gathering to fall into the wrong hands. It will be difficult to strike the right balance if the Trump administration rushes to a final decision on this matter.
The administration’s actions were set in motion in August, when the Congress enacted the Foreign Investment Risk Review Modernization Act (FIRRMA) and the Export Control Reform Act (ECRA). The former steps up scrutiny on foreign investments in the United States that could affect national security. The latter updates export controls on US technologies and products, a long overdue step.
The ECRA requires the Commerce Department to lead an interagency group to define “emerging and foundational” technologies “essential” to national security. Yet Commerce has always had the authority to unilaterally designate technologies for control. What is new is the priority and impetus given to this initiative, which may increase even further if a bipartisan bill recently introduced in the US Congress creates a new White House Office for Critical Technologies and Security to coordinate policy on these issues. Importantly, Commerce is also responsible for proposing that any controls on emerging technology also be added to the control lists of multilateral export regimes like the Wassenaar Arrangement, which includes 41 other countries. Commerce must subject controls to further scrutiny if the multilateral regimes do not adopt US controls, as controls adopted only by the United States are less likely to be effective at preventing technology leakage.
“Technology” in the export control context has a different, softer meaning than usual. It does not include finished products or “items” like software and hardware. Rather, it is information and know-how necessary to produce finished products. Therefore, the definitions and eventual lists may be more likely to influence how US firms structure their R&D operations than where they ship products.
In an inherently ambiguous situation, the administration’s notice provides little information about Commerce’s current thinking. The expansive lists in the document are only “general categories” within which Commerce “seeks to determine whether there are specific emerging technologies that are essential to…national security,” which may imply that Commerce is not aiming for broad controls that would swamp it with requests for licenses. Commerce appears to be at the drawing board, asking for input from industry on every stage of the process, from how to define emerging technology and apply it to specific technologies to how far new technologies should be developed before they are considered for controls.
But this apparent openness carries risks. While seemingly receptive to the concerns and suggestions of experts in these technologies, rather than leaving the determination solely to bureaucrats in Washington, the negotiations between defense hawks advocating wide-reaching controls and others aiming for a narrower approach are taking place in a black box. There is little clarity on the specific national security concerns that the controls should address, how Commerce generated its list of technology categories (some of which, like semiconductors for artificial intelligence, have been around since the 1990s), or what sources it uses now to keep up with emerging technologies.
Industry input is key if the law is to have well-enforced limits on which controls can be imposed. Commenters need to help Commerce figure out which controls would be useless if the technology is already available from competitors abroad. These assessments are difficult in light of the tight lid that industry keeps on this information. Indeed, companies are already reticent to go public in providing exactly the kind of proprietary, sensitive information Commerce needs to design controls effectively. Revealing this information may also have the perverse effect of helping China target its legal technology acquisition efforts before controls can be imposed and its clandestine efforts afterwards. Companies and researchers may also be reluctant to give information about their supply chains and foreign nationals involved in their technology development, even confidentially, to a US administration that may well use that information against them in future trade disputes or attempts to bring production back to the United States.
It is not clear that authorities have thought through all the unintended consequences of these efforts. For example, sharing a controlled technology with a Chinese employee of Google in California would be considered a “deemed export” with compliance requirements similar to what would be required if the technology was being shipped to Beijing instead of developed with a multinational team in Mountain View. This aspect of the controls could be abused as a backdoor way to kick Chinese employees out of US technology firms, causing the United States to fall behind other countries eager to embrace this talent.
Another problem is illustrated by a pilot program from Treasury that requires all foreigners who invest in any US company broadly involved in any technology that Commerce lists as emerging to declare their investments to the interagency Committee on Foreign Investment in the United States (CFIUS). The increasingly fluid, international process of creating technology, with major technology companies using global code bases developed by engineers and programmers located across the world, will further complicate attempts by the US government to draw borders around these technologies.
Given these complexities, the administration’s initial one-month comment period was extremely short, an indication that hawks are pushing for a quick fix. The challenge China’s industrial policy poses is real, but the unintended consequences of overly fast action are as well. The deadline was extended to January 10, but the effective time to prepare is reduced because major holidays fell inside the window. (The effect of the government shutdown makes the timetable even more unreasonable.) There will be further opportunities for industry feedback, but Commerce is under pressure to hurriedly impose controls to head off China and tailor them later, when the effects on industry emerge.
Beyond the technologies that make the list, the geographic scope of the new controls may be the most important open question. At a minimum, ECRA requires Commerce to require licenses to export to countries under a US embargo, like China and Iran. However, controls could go far beyond to hit American allies as well. Commerce’s existing process to identify emerging technologies for controls requires licenses for export to any country except Canada, but the impact has been muted because the only technology on the list as of January 2018 was related to nuclear materials.
Export controls on technology must be updated, but if the list is overly broad and lacks a targeted country focus, it could force US-based multinational firms to build walls around the United States and throw some of their foreign workforce outside them, without effectively stopping the spread of these technologies. This would disrupt networks of research and development, making everyone worse off and hastening the relative decline of US technology.
1. Section 1758(a)(2)(B).
2. ECRA Section 1758 (b)(2)(C).
3. List 0Y521, specifically 0E521.