⭐️ New Report ⭐️
What do over 2,000 open-source procurement documents reveal about China’s military robotics ambitions?
@SamBresnick takes a deep dive and outlines the PLA's pursuit of robotics across the ground, maritime, and space domains. cset.georgetown.edu/publicat…
.@SamBresnick told @politico that the effect of the new rule on remote #AI chip access “is going to be pretty immediate because you’re going to have companies outside of China violating laws if they allow Chinese companies to access the chips.” politico.com/news/2026/11/16…
⭐️ New Report ⭐️
Despite investment, hundreds of thousands of cyber job openings remain unfilled, reinforcing the belief that the U.S. faces a severe cybersecurity talent shortage.
@ali_craw and @oschinski's findings suggest a more nuanced reality ⬇️ cset.georgetown.edu/publicat…
🔷 Panel | 11:10 AM-12:10 PM
What techno-industrial threats from China is Washington missing? New fronts are taking shape in biopharma, commercial aviation, & robotics.
1⃣ Caitlin Frazer, @biotech_gov
2⃣ @sjezell, ITIF
3⃣ William Hannas, @CSETGeorgetown
4⃣ Mina Kim (moderator)
✨New Blog✨ The Nvidia scandal raises legitimate questions about China's efforts to acquire strategically important American tech as well as the responsibilities of U.S. companies that operate in commerce and national security. But the controversy also spotlights a broader danger: The United States, in responding to genuine security threats and other #AI concerns, could inadvertently undermine its own technological leadership. cset.georgetown.edu/article/…
.@SamBresnick and @HannaDohmen emphasize that there will be strong legislative pushes for tighter export controls on #AI chips following the Nvidia scandal. They state in @barronsonline that “the days of frictionless exports to Southeast Asia are likely coming to an end.” barrons.com/articles/nvidia-…
Kill switches are becoming more central to the AI safety and security conversation in Congress and it’s clear that there will continue to be AI kill switch legislative proposals in 2027 in both chambers.
Here’s a great - and brief! - @CSETGeorgetown primer on how these kill switches, once an order is issued by the government or other empowered entity, would actually work and what their effect could be.
✨New Substack Post✨ What exactly is an AI kill switch, and how would one work in practice? In this piece, we break down the concept and share key takeaways for policymakers.
Did you know CSET’s on Substack? Check out this new piece (and subscribe while you’re at it): cset.substack.com/p/ai-kill-…
✨New Substack Post✨ What exactly is an AI kill switch, and how would one work in practice? In this piece, we break down the concept and share key takeaways for policymakers.
Did you know CSET’s on Substack? Check out this new piece (and subscribe while you’re at it): cset.substack.com/p/ai-kill-…
✨New Substack Post✨ Allegations involving Nvidia chief executive Jensen Huang have renewed scrutiny of Chinese intelligence collection and U.S. controls on advanced AI chips. Meanwhile, local resistance to data centers—a sentiment that’s been linked to PRC influence campaigns—raises a different challenge to the infrastructure supporting American AI leadership. Although these episodes are not necessarily evidence of a coordinated Chinese campaign, together they illustrate how external security threats and domestic policy choices can shape the U.S.–China AI competition—and why responses must be grounded in evidence.
Read the piece on Substack (and subscribe while you’re at it): cset.substack.com/p/espionag…
From the Archive: “Through the Chat Window and Into the Real World”
In October 2024, @CSETGeorgetown published a report, co-authored by colleagues from LawAI, the University of Cambridge, METR, and others, summarizing findings from its workshop on agentic AI.
The report described four characteristics of increasingly agentic systems — complex goals, complex environments, independent planning and adaptation, and direct action — and mapped various routes for intervening to minimize risks to society. Among the routes discussed by workshop participants were legal guardrails: existing areas of law, such as criminal law, tort law, and insurance law, could play a role in how the impacts of AI agents are managed.
Two years later, AI agents have moved from “proof-of-concept products” to staple offerings from frontier AI companies — a reminder of how quickly this terrain can shift. LawAI’s Legal Frontiers team, led by Cullen O’Keefe, one of the CSET report’s co-authors, aims to incubate new law and policy proposals to address the unique challenges that accompany technological progress. Learn more about its work at docs.google.com/document/d/1…
Link to the CSET report below.
1/ 🧵 New @CSETGeorgetown report! I analyzed 2,000+ open-source PLA requests for proposal (Oct 2022–Dec 2025) to see what kinds of 🤖 robots 🤖 the Chinese military is buying.
Short version: there’s way more than humanoids.
An AI-induced human extinction scenario rests on a long chain of assumptions, each highly unlikely on its own and far less likely to align all at once. That’s not to say AI isn’t dangerous. But according to @Lauren_A_Kahn and @mchorowitz (@CFR_org), by fixating on low-likelihood, high-impact scenarios, we miss the forest for the trees: the common, mundane AI dangers already present and actively causing harm.
They urge Washington to focus on automation bias—the phenomenon where humans over-rely on or over-delegate to AI relative to the accuracy of AI models, off-loading cognition to the machine, sometimes even in the face of contradictory evidence. foreignpolicy.com/2026/11/02…
.@SamBresnick spoke with NPR’s Up First about possible national security implications of the Nvidia CEO’s relationship with an alleged Chinese spy.
Begins at 12:12: open.spotify.com/episode/1JG…
✨New Blog✨ The United States and China are pursuing two different conceptions of artificial general intelligence (AGI), but neither country has identified a definitive pathway to AGI or artificial superintelligence. American strategy must be robust to more than one possible future of AI. A few key takeaways for policymakers:
🔵 Do not mistake today’s model rankings for a permanent strategic verdict.
🔵 Strengthen the foundations of American leadership, such as research talent.
🔵 Avoid a single-track technology strategy.
🔵 Protect and cooperate selectively.
🔵 Monitor what China is actually building.
cset.georgetown.edu/article/…
1/3 @Nvidia spokespeople are (again!) claiming (falsely!) that its products are not aiding China's military modernization. We @CSETGeorgetown have a bunch of evidence that, in fact, the PLA wants to buy a range of Nvidia chips, including those that are export-controlled.