Spies Targeting U.S. AI Labs: Experts Warn Congress on China’s Espionage Push

On July 21, 2026, the House Intelligence Committee held a hearing titled “25 Years After 9/11: Confronting the Next Generation of Threats.” Experts, including former officials from the Bush and Trump administrations, highlighted the urgent need to bolster defenses for U.S. AI laboratories against foreign spying, with a primary focus on China.

Key Testimonies and Warnings:

Frank Cilluffo (Auburn University’s McCrary Institute): Stressed that China’s Ministry of State Security views frontier AI model companies as high-priority targets. He advocated rethinking counterintelligence beyond traditional “spy-vs-spy” frameworks, noting that the “keys” to advancement now reside heavily in the private sector.
Lt. Gen. H.R. McMaster (Ret., former National Security Advisor): Pointed out that post-9/11 counterterrorism focus diverted resources from China-related threats. He warned that counterintelligence has “lagged way behind,” including efforts to counter talent recruitment programs and embedded researchers in U.S. firms.
Ben Buchanan (former White House Special Advisor for AI): Noted the broader scope of modern counterintelligence, now encompassing private companies. He highlighted AI’s role in evolving both offensive cyber operations and defensive human intelligence detection. Buchanan also criticized recent policy moves easing advanced chip export restrictions to China, urging aggressive enforcement to maintain U.S. structural advantages in the AI ecosystem.
Seth Jones (CSIS): Emphasized that many U.S. states lack robust capabilities to track and counter Chinese spying activities.

The article references prior U.S. assessments of Chinese theft of military technology (e.g., aircraft data) and recent accusations of “distillation operations,” where Chinese entities covertly leverage advanced U.S. AI models to enhance their own systems.

While the White House has issued directives and quantum computing-related executive orders to protect advanced research, experts noted the absence of a comprehensive, dedicated counterintelligence framework specifically tailored to AI at the scale required.

Senate Judiciary Committee members seated in a semicircle at a hearing with a single witness at a table. — A Senate Judiciary Committee hearing with members questioning a witness seated alone.

Implications for Preparedness and National Security:

This testimony underscores the convergence of economic competition, technological superiority, and espionage in the AI domain. For veterans, preparedness communities, and private intelligence practitioners, it reinforces the need for heightened situational awareness around supply chain security, insider threats, talent protection, and emerging tech vulnerabilities. AI is not just a tool for innovation—it’s a strategic battlespace.

U.S. agencies and the private sector must accelerate collaboration to safeguard intellectual property that underpins military, economic, and societal resilience.

Overall Assessment

China is actively and aggressively pursuing U.S. proprietary AI technology through a multi-domain espionage campaign, exploiting gaps in U.S. counterintelligence posture. Frontier AI model developers and supporting private sector entities represent high-value targets because they hold the intellectual property and datasets that could enable the People’s Republic of China (PRC) to rapidly close technological gaps in military AI, autonomous systems, cyber operations, and economic competitiveness—without incurring the massive R&D costs shouldered by American innovators.

The U.S. Intelligence Community (IC) and supporting agencies have not kept pace with this evolution. Post-9/11 counterterrorism priorities diverted resources and focus away from great-power competition, leaving current counterintelligence efforts fragmented, under-resourced, and ill-suited for the private-sector-dominated AI landscape. Traditional “spy-vs-spy” models are insufficient against hybrid threats that combine cyber intrusions, human intelligence (HUMINT) recruitment/embedding, talent programs, and AI-enabled operations such as model distillation.

Confidence: High – Multiple former senior officials (including ex-National Security Advisor and White House AI advisor) provided consistent testimony aligned with prior public reporting on PRC IP theft and distillation campaigns.

Key Judgments

High Threat to Private AI Labs: PRC entities, particularly the Ministry of State Security (MSS), prioritize access to frontier AI companies. Success here would accelerate PLA modernization in areas like autonomous drones, cognitive warfare, and decision superiority.
Counterintelligence Lag: U.S. efforts remain reactive and government-centric. State-level capabilities are particularly weak, creating exploitable seams. Recent quantum computing protections offer a partial model, but no equivalent dedicated AI framework exists.
Policy Friction: Easing of advanced chip export controls undermines U.S. structural advantages in the semiconductor-AI stack. Aggressive enforcement against smuggling and circumvention is required to maintain the edge.
Broader Implications: AI espionage intersects with influence operations, supply chain vulnerabilities, and insider threats. The democratic innovation ecosystem remains America’s strongest asymmetric advantage—if properly defended.

Implications for Veterans, Communities, and Private Sector

Operational Security (OPSEC): AI researchers, developers, and supporting infrastructure face elevated risks of recruitment, phishing, and supply-chain attacks.
Preparedness: Organizations and individuals in tech, defense, and critical infrastructure should integrate counter-espionage awareness into resilience training.
Opportunity: Veteran-led OSINT and intelligence networks are well-positioned to support local awareness, threat monitoring, and bridge-building between government, industry, and community stakeholders.
Strategic Context: This is not isolated theft—it is part of systemic technology transfer strategy supporting PRC military-civil fusion.

Recommendations

Prioritize dedicated AI counterintelligence resources and public-private partnerships.
Enhance state and local awareness programs on foreign talent recruitment risks.
Maintain and strengthen export controls on enabling technologies (chips, compute, datasets).
Continue open-source monitoring of PRC AI advancements for early indicators of successful espionage.
Integrate AI domain awareness into martial arts/resilience curricula and veteran community education as a modern “situational awareness” pillar.

Source: Defense One (July 21, 2026). Analysis and adaptation by Veterans Intelligence for situational awareness.

Stay vigilant. Know the terrain.