Showcasing our researchers’ work and their latest media appearances as they weigh in on developments at the intersection of national security and emerging technology.
In an opinion piece for Foreign Affairs, Research Fellow Emily Weinstein detailed how the unprecedented response to Russia’s invasion of Ukraine has altered the culture around export controls, and how this changed environment presents an opportunity for the United States and its allies to create a new export control regime among like-minded democracies. “The nature of these challenges is shifting. So, too, must the tools that the United States and its allies use to respond to them,” Weinstein wrote. A new export control regime is needed because “none of the regimes—nor any other international body—are responsible for or designed to identify and control the various technologies that have increasingly enabled authoritarian governments to commit human rights violations.” To address current shortcomings, Weinstein recommended that the United States lead in the creation of a new regime that will “identify the commodities and technologies that currently fall through the cracks as well as the institutions and individuals in whose hands these tools could pose a threat to national security, economic security, and human rights.”
While the Bureau of Industry and Security (BIS) has added 36 new parties to the Entity List for providing support to Russia’s military in the Ukraine war, action on China’s dual-use exports lags, according to Forbes. Like Weinstein, former Assistant Secretary of Commerce for Export Administration and CSET Non-Resident Senior Fellow Kevin Wolf believes America’s export control regime is ill-fitted for today’s security challenges and needs revamping, particularly as policymakers demand levers to achieve human rights objectives. In CSET’s May 23rd webinar, A New Export Control Regime for the 21st Century, Wolf and Weinstein outlined their vision for a new export control regime among techno-democracies to better address contemporary challenges.
The Washington Post
In its first public statement on the matter, the White House has just expressed support for congressional action to require U.S. companies to disclose plans for investment in critical sectors in China. As The Washington Post reported, the U.S. government screens foreign investments in American companies that may harm national security, but doesn’t have a corresponding program to scrutinize U.S. investments in China and other countries of concern. “We really have nothing at this point in time that can deal with any outbound-investment issues,” Weinstein told The Post. Knowledge transfer is also a concern, she noted: Executives for Sequoia Capital China, an affiliate of Silicon Valley’s Sequoia Capital, sometimes sit on the boards of firms they invest in. Lending that management expertise and credibility to companies trying to establish themselves in the global market is “huge,” Weinstein said.
MIT Technology Review
Since the Russian invasion of Ukraine began, militaries have looked to push more artificial intelligence technology onto the battlefield, MIT Technology Review reported. Citing a CSET report by Ryan Fedasiuk, Jennifer Melot and Ben Murphy, Harnessed Lightning, the publication pointed out that the Chinese military likely spends at least $1.6 billion a year on AI development. However, getting militaries to adopt and implement AI technology takes time and expertise. Research Analyst Ngor Luong, who has studied AI investment trends for CSET, noted that AI companies with military ambitions have to “stay in business for a long time.”
Inside Higher Ed
AI faculty are in short supply at U.S. universities, but it’s not a result of the job market. According to a just-published CSET report, university hiring has not been able to keep pace with the high student demand for AI instruction. Inside Higher Ed reached out to Research Analyst and co-author of the report Jack Corrigan to learn how universities are responding to the AI faculty shortages. Due to increased student demand, universities have restricted access to AI programs by limiting enrollment in high-demand classes, reducing the number of small-enrollment classes and tightening computer science admission requirements, said Corrigan. The Register also highlighted the report, noting in particular its observation that “teaching capacity gaps limit the amount of talent flowing into the US AI workforce, which in turn negatively impacts economic and national security.”
In an opinion piece for The Hill, Research Analyst Luke Koslosky argued that community colleges are a massive source of potential for expanding the AI workforce. Drawing from his CSET report, co-authored with Research Fellow Diana Gehlhaus, he wrote that “community colleges could create pathways to good-paying jobs across the United States and become tools for training a new generation of AI-literate workers. ” Instead of focusing solely on four-year bachelor’s degree candidates, Koslosky encouraged AI employers to broaden and diversify their workforce to consider credentials that signal competencies. To maximize talent from community colleges, Koslosky recommended that the National Institute of Standards and Technology (NIST), or another suitable government agency, create a framework for work roles and competencies for AI jobs for schools to design curricula suitable to industry needs. “Leveraging community colleges offers a way for the United States to outpace its competition, create upward mobility for millions of workers and adapt its workforce for the jobs of the future,” Koslosky wrote. “But they need help to get there. With the right support from policymakers and buy-in from the many other stakeholders needed, they can turn their potential into reality.”
Construction of new semiconductor manufacturing facilities in the United States to keep up with demand could create as many as 27,000 new positions in the semiconductor industry according to CSET research cited in an article by Protocol. However, relying solely on U.S.-born talent to fill the positions will not be enough. Research Analyst Will Hunt, who co-authored the report with CSET colleague Owen Daniels, estimated in earlier research that at minimum, 3,500 foreign-born students — mostly engineers from Taiwan and China — will need to immigrate and get specialty visas to staff the planned facilities.
Tech Times took note of Silicon Twist, which revealed the Chinese military’s acquisition of computer chips that are developed by U.S. companies. After looking at thousands of Chinese purchase records, the report’s authors – Ryan Fedasiuk, Karson Elmgren and Ellen Lu – identified 97 such AI chips designed by U.S. companies Nvidia, Xilinx (now AMD), Intel, or Microsemi. According to the authors, “public equipment contracts indicate that the Chinese military purchases U.S.-origin equipment through intermediaries, including both officially licensed distributors and shell companies, not directly from U.S. semiconductor vendors.” To effectively manage the Chinese military’s access to AI chips, the authors recommended that the U.S. government gain a better understanding of the Chinese defense industry and impose new forms of export controls for AI technologies.
The Wire China
The Wire China reached out to Translation Manager Ben Murphy to learn more about China’s “chokepoints,” or key technological dependencies, as detailed in his recent report – and made that report the centerpiece of a story. Murphy analyzed a series of articles published by a Chinese government agency, never before translated together in English, that listed 35 areas of such dependency – some of them easy to overlook. As colleague Emily Weinstein noted, “It’s not just about protecting what’s new and innovative like AI or quantum. We also have to pay attention to the things that are frankly kind of boring, like aviation grade steel.” Understanding China’s chokepoints can offer an important source of leverage for strategic gains by the United States. “Lots of Chinese policy documents list out categories of technology that China feels it has to reduce its import dependencies in,” said Murphy. “What’s special about this series is they actually get into specifically, within those categories, what it is that [China needs].”
In a policy brief for Brookings, Research Analyst Dahlia Peterson and coauthor Samantha Hoffman outlined the use of surveillance technology and concerns regarding privacy and human rights abuses. According to the authors, “the United States and partner democracies have implemented sanctions, export controls, and investment banks to rein in the unchecked spread of surveillance technology, but the opaque nature of supply chains leaves it unclear how well these efforts are working.” To address policy challenges, the authors provided recommendations for democratic governments and civil society to better govern the use of surveillance technologies.
A lengthy excerpt from a CSET report by John VerWey, Re-Shoring Advanced Semiconductor Packaging, was repurposed as a bylined article in Semiconductor Digest’s July 2022 issue. The report detailed how targeted investment incentives to increase U.S.-based advanced packaging capacity are also important for increasing semiconductor supply chain resilience.
Spotlight on CSET Experts: Emily Weinstein
Emily Weinstein is a Research Fellow focused on U.S. national competitiveness in AI/ML technology and U.S.-China technology competition.
In her previous role at CSET, Emily conducted research on China’s S&T ecosystem, talent flows, and technology transfer issues. Emily has testified before the U.S.-China Economic and Security Review Commission and the Wisconsin State Legislature’s Senate Committee on Universities and Technical Colleges.
Her latest publications include China’s State Key Laboratory System, Headline or Trend Line?, and China is Fast Outpacing U.S. STEM PhD Growth. Her work has appeared in Foreign Affairs, Foreign Policy, The Washington Post, Lawfare, DefenseOne, and other outlets.
Interested in speaking with Emily or our other experts? Contact External Affairs Specialist Adrienne Thompson at firstname.lastname@example.org.