⚡NIMITZ TECH NEWS FLASH⚡

“Protecting Our Edge: Trade Secrets and the Global AI Arms Race”

House Judiciary Subcommittee on Courts, Intellectual Property, Artificial Intelligence, and the Internet

May 7, 2025 (recording linked here)

HEARING INFORMATION

Witnesses and Written Testimony (Linked):

• Mr. Nicholas Anderson: President and Chief Operating Officer, Invictus International Consulting, LLC

• Dr. Benjamin Jensen: Senior Fellow, Center for Strategic and International Studies

• Mr. Christopher Mohr: President, Software and Information Industry Association

• Ms. Helen Toner: Director of Strategy and Foundational Research Grants, Georgetown University Center for Security and Emerging Technology

• Dr. John Villasenor: Professor of Electrical Engineering, Law, Public Policy, and Management, University of California, Los Angeles
  

HEARING HIGHLIGHTS

IN THEIR WORDS

SUMMARY OF OPENING STATEMENTS FROM THE COMMITTEE AND SUBCOMMITTEE

• Subcommittee Chair Issa opened the hearing by highlighting the central conflict in artificial intelligence: balancing transparency with the protection of proprietary innovation. He stressed that AI training algorithms and datasets represent invaluable trade secrets that must be shielded from foreign adversaries, particularly China. Issa criticized the idea of mandatory disclosures under the guise of transparency, warning such moves could stifle innovation and benefit hostile nations. He emphasized the importance of bipartisan efforts to accelerate legal protections and prevent intellectual property theft.

• Subcommittee Ranking Member Johnson expressed concern over the rapid development of Chinese AI models like DeepSeek, suggesting they may have violated IP rules to leap ahead. He cautioned against sacrificing responsible innovation for speed, likening reckless AI growth to fragile plants that grow fast but are prone to collapse. Johnson called for robust cybersecurity standards, IP protections, and support for international collaboration, while rejecting isolationist or anti-immigrant policies that damage innovation. He warned that policies under the Trump administration, such as visa revocations and unpredictable tariffs, have harmed American AI leadership.

• Full Committee Ranking Member Raskin echoed bipartisan support for the U.S. intellectual property system but criticized the Trump administration for undermining it with erratic trade wars and attacks on universities. He emphasized that instability discourages investment and creativity, especially from independent inventors and small businesses. Raskin argued that innovation depends on a functioning and fair judicial system to enforce IP rights. He concluded that while AI competition is inevitable, the theft of American innovation by foreign actors like China demands serious attention.

SUMMARY OF WITNESS STATEMENT

• Dr. Jensen framed AI as a defining global competition and emphasized the geopolitical stakes of agentic AI, which integrates systems to influence decision-making at scale. He warned that China is using state-backed strategies to steal U.S. AI innovations through knowledge distillation, cyber-espionage, and regulatory loopholes. Jensen called for modernizing IP law, classifying AI companies as critical infrastructure, and enhancing offensive cyber capabilities to deter theft. He stressed that preserving American leadership in AI is about defending free societies against authoritarian control.

• Dr. Villasenor affirmed that the U.S. leads in AI but warned that innovation makes American companies prime targets for trade secret theft. He criticized expansive transparency mandates that could inadvertently expose AI model security details, citing a revoked executive order as a cautionary example. He raised novel legal questions about asserting trade secrets in complex, self-adapting AI systems, emphasizing the need for clear standards. Villasenor opposed overregulation, arguing it could cripple U.S. AI leadership and empower geopolitical rivals like China.

• Mr. Mohr warned that the PRC remains deeply involved in state-backed IP theft. He described evolving threats to trade secrets, especially in the age of AI-as-a-service, and noted that bad actors aim to reverse-engineer innovations. Mohr proposed stronger cybersecurity support for smaller firms, broader CFIUS oversight, cooperation with allies on secure AI infrastructure, and scrutiny of foreign-funded litigation exploiting the U.S. legal system.

• Ms. Toner argued that security and transparency can be balanced, with distinct categories of data requiring protection versus those needing oversight. Toner recommended expanding public-private security collaborations, mandating transparency, investing in secure infrastructure, and legally enabling risk-reducing practices like personnel vetting. She emphasized that failing to secure these systems risks handing strategic advantage to foreign adversaries.

• Mr. Anderson warned that China seeks to dominate AI through theft, espionage, and deceptive investment tactics. He criticized the current lack of mandatory security standards, outbound investment vetting, and transparency in AI firms’ foreign ownership. Anderson called for policies that impose discipline and prioritize national security, including stronger cyber requirements and more aggressive counterintelligence operations.

SUMMARY OF KEY Q&A

• Rep. Fitzgerald asked whether AI should be considered critical infrastructure. Mr. Anderson agreed, saying it would allow the government to set minimum cybersecurity standards and integrate AI firms into broader national security planning through agencies like DHS and CISA.

  Rep. Fitzgerald asked how trade secrets are protected in litigation. Mr. Mohr explained that judges typically issue protective orders that restrict access to confidential material to prevent leaks, since once a trade secret is exposed publicly, it loses its value. Rep. Fitzgerald asked how foreign-funded litigation increases trade secret risk. Mr. Mohr said protective orders can vary based on the nature of the case; if the funding source is a hostile foreign entity, judges may impose stronger restrictions and personal liability on attorneys to prevent misuse.

  Rep. Fitzgerald asked why trade secrets are especially important in AI. Dr. Villasenor explained that even open models contain proprietary data, training methods, and code that are rarely shared. AI’s scale and complexity make trade secrets the main mechanism of competitive protection.

• Ranking Member Johnson inquired about whether avoiding AI regulation to outpace China is a valid strategy and what advantages the U.S. holds. Ms. Toner argued that smart regulation can build public trust, enhance safety, and support U.S. industry without slowing innovation. She said compute access is the U.S.’s biggest advantage and export controls are working. Overly rigid checklists should be avoided, but transparency requirements can help without harming competitiveness.

  Ranking Member Johnson asked whether all transparency obligations threaten trade secrets. Dr. Villasenor said many do, but disclosure rules can be tailored by sector to minimize the risk while maintaining accountability.

• Rep. Massie asked what government action is most important and which direction of information sharing matters more. Ms. Toner recommended expanding voluntary partnerships between AI companies and national security agencies, especially around threat intelligence sharing and collaborative testing. She said both are critical, but government-to-industry sharing is especially valuable given access to classified data.

  Rep. Massie brought up whether AI collaboration should be mandatory. Dr. Jensen said voluntary approaches are best but supported minimum cybersecurity standards. He warned against burdensome regulation that might shift company focus away from innovation.

  Rep. Massie asked if trade secrets are preferred over patents for AI. Dr. Villasenor explained trade secrets cover far more than patents, especially for source code, data, and system behavior not eligible for patent protection. He also acknowledged insider hiring as a major risk alongside hacking and espionage, though companies must rely on employee integrity.

• Rep. Raskin inquired how open-source components affect trade secret protections and whether enforcement remains viable. Dr. Villasenor explained that trade secrets can remain valid if proprietary elements are added to open code. Dr. Jensen noted enforcement alone is inadequate and urged legal and intelligence strategies to counter foreign IP theft.

  Rep. Raskin explored transparency should entail in high-stakes AI. Ms. Toner recommended layered transparency: public disclosures for accountability, confidential reporting to government, and third-party audits under NDAs to verify safety claims.

  Rep. Raskin pressed on immigration limits affecting the AI workforce and highlighted the role of civil society in model oversight. Dr. Jensen warned that such limits would cripple labs, including his own, and emphasized either reform or heavy investment in U.S. STEM education.He also supported independent nonprofits for benchmarking AI behaviors, especially in sensitive domains like foreign policy modeling.

• Rep. Fry probed the connection between AI leadership and national defense. Dr. Jensen affirmed its criticality, describing successful experiments using AI in Marine Corps planning, and emphasized the need to train military personnel in AI systems. Rep. Fry sought clarity on Chinese espionage’s impact on defense and asked about China’s recruitment of U.S. researchers through the Thousand Talents Plan. Dr. Jensen said it erodes innovation by reducing China’s R&D costs and exposing U.S. military patterns. He advocated for countermeasures like data deception and algorithmic decoys. Ms. Toner confirmed its focus on AI experts and said China exploits U.S. visa uncertainty to lure talent with lucrative research roles. Rep. Fry followed up on whether these efforts have compromised IP. Ms. Toner said yes, particularly when employees take knowledge from U.S. firms to Chinese institutions, and called for stronger vetting of sensitive roles—while cautioning against anti-immigrant policies.

• Rep. Ross questioned whether DeepSeek’s AI model resulted from theft or legitimate innovation. Dr. Jensen described it as legal gray-area theft via model distillation and API manipulation but said its biases could become strategic liabilities. Ms. Toner countered that DeepSeek employed standard industry practices and meaningful engineering, warning against underestimating their capabilities.

  Rep. Ross asked how to measure authentic R&D and raised concerns about cutting NSF grants tied to DEI. Dr. Jensen pointed to patent filings, talent migration, and public research investments as key indicators. He also cautioned such moves to cut NSF grants could undercut inclusion and weaken the diversity essential to long-term innovation.

• Rep. Cline sought insight on the role of trade secrets in AI competitiveness and how to prevent IP theft from foreign actors. Mr. Anderson called them vital, comparing cybersecurity protections to combat armor in hostile environments. He recommended cybersecurity hygiene, employee vetting, and stricter controls on outbound investments. Rep. Cline explored how to balance innovation with AI transparency. Mr. Anderson proposed setting clear redlines for what must be disclosed, especially for high-risk systems, without undermining proprietary value. Rep. Cline asked if existing laws offer a framework and requested details on critical areas for trade secret protection. Mr. Anderson endorsed CISA 2015 as a working model for public-private info-sharing and supported its reauthorization. Mr. Mohr emphasized the need to protect algorithms, training data, chip configurations, and opaque system behavior—particularly where current law may not suffice.

• Rep. Kamlager-Dove pressed on the risks of restricting immigration for AI talent and stressed the value of immigrant contributions. Ms. Toner said attracting global researchers is the U.S.’s greatest edge over China and essential to remain competitive given China’s vast STEM pipeline. She noted immigrants comprise over half of U.S. AI PhDs and founded two-thirds of leading AI startups.

  Rep. Kamlager-Dove sought long-term implications of driving talent away. Dr. Villasenor warned that displacing immigrant researchers would harm future job creation and innovation, as they often become startup founders and major employers.

• Chair Issa challenged the assumption that Chinese AI students return home. Ms. Toner clarified that over 85% stay long-term, though many rely on complicated visa routes without a guaranteed path to permanence.

  Chair Issa dismissed tariffs as a tool for protecting AI leadership. All panelists concurred that tariffs are ineffective, while export controls—especially on chips—have proven more valuable.

  Chair Issa suggested tying access to restricted chips to national cooperation. Dr. Jensen and Ms. Toner agreed this could drive better security practices and incentivize collaboration with U.S. agencies.

  Chair Issa raised concerns about AI-as-a-service undermining export controls. Dr. Jensen confirmed that output harvesting and API manipulation let adversaries sidestep hardware restrictions. He called for aggressive intelligence declassification and legal coordination. Ms. Toner stressed the need for AI firms to monitor system usage closely and implement robust audit processes to detect misuse.

ADD TO THE NIMITZ NETWORK

Know someone else who would enjoy our updates? Feel free to forward them this email and have them subscribe here.