National Security Data and Policy Institute

Cryptocurrency exchange networks have provided the means for crimes like money laundering and narcotics trafficking; their anonymity and strategically deceptive actors make identifying illicit financial activities a major challenge. Deploying advanced analytical tools that use machine learning can help stop these crimes, but training these tools is complex and expensive.     How can financial intelligence and law enforcement agencies address this growing threat with finite resources and a scarcity of expertise?     This NSDPI research paper ( 🔻Research Linked Below 🔻 ) offers practical approaches to balance automated detection with hands-on expert analysis of illicit financial activity, exploring a leaner and smarter way to detect illicit crypto transactions with well-trained machine learning models. We explore optimized methods for labeling data for ML models and show how agencies can minimize the cost and expert oversight required to effectively train these AI tools. Here are some high-level takeaways:    🔶 Using data labeling to identify only a small fraction of illicit nodes (< 4%) is enough for supervised machine learning models to outperform unsupervised models.   🔶 Integrating active learning with other cost-sensitive labeling techniques can optimize intelligence resources in crypto monitoring, pointing to a scalable solution that may generalize to other contexts.   🔶 Employing Graph Neural Networks (GNNs) as an unsupervised method to detect anomalies only requires sampled subgraphs for training, significantly lowering memory costs.    To learn more about these methods for effectively training machine learning tools to find illicit financial activity, read the full paper: https://lnkd.in/ehiBbV7z #AI #cryptocurrency #MachineLearning

Microelectronics are essential for critical technology like defense systems, clean energy, and robotics, but they depend on a complex network of global trade. Any disruption to these supply chains is a serious threat to global security and industry. But how can policymakers untangle complicated trade patterns to identify points of systemic vulnerability and mitigate national risk?  To better approach this question, the NSDPI analyzed microelectronic supply chains using an algorithm created with network analysis, machine learning, and qualitative analysis ( 🔻 Paper Linked Below 🔻 ). By applying this algorithm to UN Comtrade trade data from 2017 to 2024, we rank countries whose removal from the world’s microelectronics trading system would create the biggest shock, and show how risk exposure converges or differs across the United States and China.  Our algorithmic analysis uncovers key microelectronics trade patterns and vulnerabilities, including:  🔶 The U.S. is disproportionately exposed to disruption by exporters of processed inputs, especially from China, the UK, Germany, and the rest of the EU.  🔶 While China’s supplier dependency is generally less than the U.S., China is highly exposed to and reliant upon exports from Brazil, Myanmar, and Thailand.  🔶 Brazil consistently emerges as an exporter that both the U.S. and China are highly dependent upon.  To learn more about the unique findings our algorithm uncovered about #microelectronics #supplychains read the full white paper: https://lnkd.in/e7YkHif3

China will be the world’s leading center of #AI innovation by 2030, according to public announcements by #PRC leadership. Chinese government documents can be prone to hyperbole, but our researchers at the NSDPI found China’s socio-institutional culture is setting this ambition up to succeed.  Our recent white paper (🔻Link Below🔻) explores why, in the words of Nvidia CEO Jensen Huang, “China is not behind anybody” in AI development. Key takeaways include: 🔶 Why the NGAIDP’s extensive detail into local implementation of its national AI plan and its roadmap for equitable distribution of AI support indicate China is serious about cultivating AI leadership across the country—even as it provides special investment in cities where AI companies are already thriving.     🔶 How a 2017 GO Match between Chinese grandmaster Ke Jie and Google’s DeepMind AI platform and its aftermath revealed China’s government and citizens view the disruptive nature of AI more positively than many nations.      🔶 How PRC support and funding have already led to Chinese companies emerging as global AI leaders, from Chinese company DeepSeek’s efficient, cost-effective, and open-source AI model to China filing 61.5% of 45,000 AI-related patents published in 2023.   Get the full story of how #China sees AI as a disruptive technology that links with Beijing’s sweeping vision of sociopolitical security by reading our full paper: https://lnkd.in/e62WPHWx

ICYMI, some of our interns shared their research on the airwaves last week at Charlottesville’s local NPR affiliate!    Eli Beaubien, Cole Leis, and Robert James May joined Radio IQ/WVTF Public Radio’s Sandy Hausman to talk about their ten-week internships spent studying issues of interest to UVA’s National Security Data and Policy Institute (NSDPI). Their final white papers explored the challenges around protecting American investment in #AI, mining Helium-3 from the moon, and safeguarding against increasingly sophisticated deepfake technology. Click below to listen to the full piece! https://lnkd.in/ePaJV7ZQ Phil Potter George Foresman Jonathon Hathaway John Robinson University of Virginia Frank Batten School of Leadership and Public Policy

We're thrilled to announce the formation of our inaugural National Security Data and Policy Institute (NSDPI) Policy Advisory Group (PAG)!   The 2025 PAG is composed of some of the top minds in #AI automation, #cybersecurity, tech investment, governmental advisory, academia, and other fields. With its expertise, this body is uniquely positioned to help the Institute frame its research and planning going forward. These leaders will convene at least twice a year to review reports on the NSDPI by its executive director, receive updates on some of the Institute’s key programs and projects, and engage in robust discussions of strategy. This development comes at an exciting time for the Institute, as we celebrate a full year since we were established in partnership with the Office of the Director of National Intelligence in August of 2024. The NSDPI has served as a unique hub for innovation since then, bringing together academia, government, and industry to address critical issues in national security. These partnerships have produced a wealth of publicly available research in data science, artificial intelligence, research computing, policy, and other fields; with the benefit of the PAG’s combined decades of expertise, we’ll continue to expand on this important work.    Read More about the PAG:  https://lnkd.in/e_iecwmw   Browse the NSDPI’s Library of Research:  https://lnkd.in/e9Waae9m

The white papers and literature reviews these students produced will serve as a valuable resource for our researchers as they continue to delve into some of the most pressing national security issues facing the U.S. Our thanks to the NSPC's 2025 summer interns for their excellent work and dedication!

“We’re most vulnerable if we lose sight of our own system and what we’re good at... the ability to foster a system that produces innovation and flexibility in areas where maybe we aren’t anticipating the need for flexibility is the right way to go. It’s hard to call your shots in advance if you don’t know where you’re going or when you need to get there.”     NSDPI Executive Director Phil Potter gave his thoughts on the future of #AI competition between the US and the #PRC as part of a recent panel discussion hosted by the team at Raft. Sitting down for an interview with Breaking Defense afterward, he shared more on the direction he sees AI innovation taking and the important role the NSDPI plays in that framework. Our thanks to Raft for hosting this expert exchange of ideas, and to Breaking Defense for bringing this issue to the forefront.     🔎 The NSDPI has produced new, in-depth #Research on the social foundations of China's AI #Policy. Click here to read our key takeaways, download our work, and explore related topics in #NationalSecurity: https://lnkd.in/e62WPHWx 🎥 Click below to watch Director Potter's full Breaking Defense interview:

We received outstanding work last week from these emerging leaders at the National Security Policy Center--we look forward to more insights from this talented group in the last weeks of their internship.