Just ahead of a holiday break, our experts are wrapping up the year by weighing in on a variety of national security and emerging technology issues in the media. This month, they covered AI techniques in cyber defense, the effects of China’s use of state-sponsored espionage, U.S.-China tech competition, and more.
The New York Times
Increased scrutiny of Chinese academics and scientists by the U.S. Department of Justice’s China Initiative has caused concern and disruption at many universities, according to a New York Times article. “There’s no room for xenophobia or ethnic profiling,” said CSET Director of Biotechnology and Senior Fellow Anna Puglisi. “But what gets lost in the discussion is the bigger question that we need to be asking, which is: ‘Do we have the system in place to mitigate the behavior and central government policies of a nation state that are specifically set up to target the seams in our system?'” Puglisi discussed this subject previously in her testimony before the Senate Select Committee on Intelligence, where she offered recommendations to counter China’s S&T collection efforts.
Marketplace public radio reached out to Puglisi about China’s systematic use of state-sponsored industrial espionage. “China really looks at development of science and technology as zero-sum. That’s really the driver behind a lot of the activities that we see,” she said. Drawing from a recent report on the subject that she coauthored, Puglisi discussed how “nonmarket decision-making and state subsidies give unfair advantage to China’s companies and force U.S. and other Western companies to have to make concessions and give up technology they do not have to do other places in the world.”
In an opinion piece for Foreign Affairs, CyberAI Director John Bansemer discussed how AI can be leveraged for cyberattacks while also bolstering cyber defense. If the United States is to prepare defensive AI capabilities, he wrote, researchers need realistic data, and the federal government must prioritize AI-enabled research and development. “The impact of uncontrolled cyberattacks is becoming ever more costly,” Bansemer asserted. “It is past time for the United States to explore the potential for AI to improve its cyber defenses to better protect critical infrastructure providers and state and local governments.”
In an opinion piece for Foreign Policy on U.S.-China competition, Research Analysts Ryan Fedasiuk and Emily Weinstein laid out four key elements at the core of China’s strategy to become a technology superpower. “These four elements—equipment, personnel, information, and capital—are the four pillars of China’s playbook to become a technology superpower,” they wrote. “Each of these pillars presents a challenge to the United States—but also offers leverage points and opportunities for America to defend and sharpen its technological edge.”
The Washington Post
Fedasiuk’s latest report was featured in a Washington Post story on U.S. government concerns about investments in Chinese companies with military or surveillance ties. The report, which Fedasiuk wrote with Jennifer Melot and Ben Murphy, identifies China’s key AI defense industry suppliers, highlights gaps in U.S. export control policies, and contextualizes the People’s Liberation Army’s AI investments within China’s broader strategy to compete militarily with the United States. “It’s clear there are large gaps in the U.S. export control system that allow the Chinese military to access equipment, information and capital originating in the United States,” Fedasiuk told The Post. “But plugging these gaps is easier said than done.”
Will China’s regulatory “Great Wall” hamper AI ambitions? In The Diplomat, Fedasiuk and Ellen Lu examine how China’s Data Security Law (DSL) and Personal Information Protection Law (PIPL) will affect state security bureaus’ data collection capabilities, and create steep compliance costs for internet companies that could hamper China’s ambitions to be the leader in AI by 2030. The authors argue that “the DSL and PIPL will make it substantially harder for existing businesses to continue operating with the same degree of autonomy they had enjoyed in the past, and may create steep barriers for new players hoping to enter China’s tech market.”
Director Dewey Murdick was asked to join the ChinaTalk podcast to discuss CSET’s mission and explore a range of topics from semiconductors to biotechnology. In the interview, Murdick explained the research behind CSET’s Map of Science. “Every problem we pick on or every solution that we work toward building at CSET is motivated by a policy-relevant problem. … And because we’re dealing with emerging technologies, the research literature is actually relevant. Now we’re working to actually connect that research literature to future literature,” Murdick said.
Voice of America
Voice of America spoke with CSET Research Analyst Ngor Luong about the United States government’s plans to advance its strategic partnership with the Association of Southeast Asia Nations (ASEAN) through an Indo Pacific Economic Framework. However, the Framework is not just about trade, according to Luong. “Take semiconductors for example; as significant players in chip packaging and manufacturing, ASEAN countries such as Malaysia and Singapore may play an important role in helping the United States build resilient supply chains for semiconductors.” Amid China’s aggression in the South China Sea, the United States will have to decipher mixed signals from ASEAN members. “China’s actions and words don’t typically align, and the challenge facing ASEAN countries is to maintain their independence from China’s sphere of influence,” said Luong.
Director of Data Science and Research Catherine Aiken outlined why classifying AI systems is necessary and explained the frameworks involved in her research in partnership with OECD.AI. In addition to her recent report, CSET also launched an interactive website to explore the different frameworks.
A new report by Tim Rudner and Helen Toner introduces specification in machine learning as a key element in greater AI safety. The EE Times explored the key challenges behind the use of specification in machine learning. The authors noted that “as machine learning systems become more advanced, they will likely be deployed in increasingly complex environments to carry out increasingly complex tasks. … This is where specification problems may begin to bite.” Previously, Toner was featured in an EE Times podcast discussing what safe and reliable AI looks like.
Spotlight on CSET Experts: Anna Puglisi
Anna Puglisi is CSET’s Director of Biotechnology Programs and Senior Fellow. She has previously served as the National Counterintelligence Officer for East Asia and played a prominent role in drafting the U.S. National Counterintelligence Strategy.
Her latest reports include The Huawei Moment and China’s Foreign Technology Wish List. She has been featured and quoted on her expertise in counterintelligence and China in a variety of outlets, including Reuters, The Washington Post, The Wall Street Journal, and National Public Radio.
Interested in speaking with Anna or our other experts? Contact External Affairs Specialist Adrienne Thompson at firstname.lastname@example.org.
Want to stay up to date on the latest CSET research? Sign up for our day-of-release reports and take a look at our biweekly newsletter, policy.ai.