Today, I testified before the Office of the United States Trade Representative. 

Click here to download a PDF of the written testimony

Thank you for the opportunity to testify today. I am Evan Swarztrauber, a senior fellow at the Foundation for American Innovation, a technology policy think tank based here in Washington. My organization believes that technology should serve humanity, individual freedom, and U.S. national security. To that end, I testify in support of USTR’s Section 301 investigation into China’s (the Chinese government) predatory subsidies and other market distortions in the semiconductor industry. USTR should take decisive action to investigate and address China’s aggressive, state-backed efforts to dominate the global silicon carbide wafer (SiC) market. These wafers are key inputs in a wide range of critical areas, including electric vehicles, power grids, radar systems, and missile defense systems. The U.S. must not lose its capacity to produce SiC domestically.

This should be obvious, but it bears repeating: China does not subsidize industries out of generosity. They do so with a clear and deliberate strategy: to undercut foreign competitors, monopolize critical supply chains, and gain leverage to support the economic and security goals of the Chinese Communist Party. We have seen this playbook before—whether with solar panels, LED screens, or lithium-ion batteries. Now, China is deploying the same tactics in the SiC wafer sector. Through massive government subsidies, below-cost pricing, and market manipulation, Chinese firms are working to drive American and allied producers out of business, leaving the U.S. and its partners dependent on Chinese supply. In my view, this behavior is unreasonable, discriminatory, and should merit an actionable determination by USTR.

For too long, the U.S. government had been largely content to let foreign adversaries flood our markets with artificially cheap goods. U.S. trade policy had prioritized short-term savings and consumerism over long-term economic and national security. As a result, Americans watched entire industries and communities collapse.

The outdated approach ended when President Trump was first elected in 2016, and there is now substantial bipartisan support for protecting high-tech American manufacturing. The CHIPS Act represents a major effort to atone for past sins, which is especially important given China’s increasing aggression toward Taiwan and elsewhere in the Pacific. Yet, unsurprisingly, China is taking steps to undermine our progress, specifically by flooding the market with heavily subsidized SiC wafers.

These wafers enable high-performance electronics used in electric vehicles, power grids, telecommunications, and advanced military systems. Without a secure and competitive domestic supply, the United States risks falling behind in these strategic industries, leaving us vulnerable to supply disruptions and economic coercion, especially in the event of war.

USTR should conduct a thorough and robust investigation into China’s actions in the SiC market and take any and all necessary measures to counteract these distortions. America’s blanket 20% tariff on China is welcome but insufficient to correct for China’s excess subsidies and distortions on particular products like SiC. USTR should consider a significantly higher and targeted tariff for these PRC-produced wafers as well as final products that use PRC-produced SiC.

Second, USTR should investigate how Chinese state-owned and state-affiliated firms use third countries as a loophole to evade U.S. trade restrictions and hide their ownership interests. While more facts are yet to come, the advancement of the Chinese AI startup DeepSeek is a potential example of how China uses shell games in Singapore and other nations to avoid sanctions, exploit U.S. technology, and steal intellectual property.

Third, the U.S. government should use its purchasing power to prioritize domestically produced SiC wafers and semiconductors, especially for military purposes. There is no reason that American taxpayers should be subsidizing a Chinese rope to hang us, when they could instead help build an American life raft for SiC production.

When it comes to China’s trade practices, we may have learned our lesson. But learning a lesson is no use if not implemented. We must act decisively to protect our capacity to build SiC domestically both for current use cases and those we cannot yet imagine. Thank you for your leadership and public service on this vital issue, and I welcome any questions.

Section 301 Investigation on China’s Acts, Policies, and Practices Related to Targeting of the  Semiconductor Industry for Dominance

How secure is America’s lead in artificial intelligence? It wasn’t long ago that the conventional wisdom put China years behind the curve. This was partly because the United States introduced export controls in 2022 and 2023 that notionally embargoed Chinese companies from the most advanced AI chips.

Today, however, most experts I speak with believe China trails the U.S. on AI by at most six to nine months—if at all. Take the sudden rise of China’s DeepSeek, the impressive AI startup that began as a hedge-fund CEO’s side project. DeepSeek shocked the markets earlier this year by releasing R1, a “reasoning model” that replicated top AI firm OpenAI’s breakthrough o1 model only a month after the latter’s unveiling, and seemingly at a fraction of the cost.

OpenAI’s latest models still top the leaderboards, but only because they have access to vastly more computing resources. What DeepSeek nonetheless demonstrated is just how few technical barriers stand in the way of competing at the AI frontier. As DeepSeek’s CEO Liang Wenfeng put it 

 last year, “money has never been the problem for us; bans on shipments of advanced chips are the problem.”

U.S. Companies Are Helping China Win the AI Race

Today, I submitted a comment in response to the National Science Foundation and Networking and Information Technology Research and Development Program’s request for information regarding the development of an Artificial Intelligence Action Plan. Click 

Thank you for the opportunity to comment on the National Science Foundation’s (NSF) and the Network and Information Technology Research Development’s (NITRD) request for information on the development of the Artificial Intelligence (AI) Action Plan. In my comment, I seek to draw attention to an issue that significantly impacts AI development: copyright law. This comment is adapted from a recent paper I co-authored with Tim Hwang, titled: “Copyright AI, and Great Power Competition.”

This document is approved for public dissemination. The document contains no business-proprietary or confidential 

information. Document contents may be reused by the government in developing the AI Action Plan and associated documents without attribution.

Artificial intelligence (AI) has the makings of a general-purpose technology (GPT) that could usher in a new industrial revolution. As a GPT, AI can be diffused into existing processes while creating new products, services, and markets. This has significant implications for national soft power (economic strength, technological adoption, cultural impact) and hard power (military capabilities, statecraft, national sovereignty, and geopolitical competition).

The most significant geopolitical AI competition is between the United States and China. The Chinese Communist Party (CCP) has made clear its desire to change the status quo, with AI featuring heavily in such plans. In 2018, President Xi Jinping stated that China must "ensure that our country marches in the front ranks where it comes to theoretical research in this important area of AI, and occupies the high ground in critical and AI core technologies." In 2023, Xi directed the People's Liberation Army to "raise the presence of combat forces in new domains and of new qualities."

2. Copyright Law as a Threat to U.S. AI Leadership

There is one area of law that could deal a serious blow to U.S. efforts to maintain AI leadership: copyright. Nearly all leading AI model developers in the U.S. face lawsuits from copyright holders claiming that the inclusion of copyrighted material in training materials constitutes infringement. Currently, 39 open copyright cases against model developers seek damages and ask for data sets to be destroyed and models retrained. If plaintiffs are successful, they would cause irreparable harm to America's AI sector, with grave implications for America's economy, military capabilities, and geopolitical standing. Identifying the threat copyright lawsuits pose to American model developers and operationalizing policies to mitigate potential harms will be critical to advancing American competitiveness in AI development.

Current "frontier" AI models would not exist without the ability to access and analyze vast amounts of data available online. Training large-scale models using prevailing scaling techniques requires massive data sets that can provide the necessary diversity, complexity, and quality of data. The default rule under U.S. copyright law is that unauthorized copying of material is actionable infringement. However, AI firms argue that accessing and analyzing such data constitutes fair use.

Fair use is a doctrine intertwined with copyright law for centuries and officially enshrined in the Copyright Act of 1976. It allows for nonpermissive use of copyrighted works in a transformative manner as a counterbalance to the limited monopoly that copyright holders enjoy. It is seen as a counterbalance to the limited monopoly that copyright-holders enjoy over their work to “promote the Progress of Science and useful Arts,” the original justification for copyright protection. Fair use is determined on a case-by-case basis, with judges relying on a four-part test to determine whether the use of a copyrighted work qualifies as fair use. The four aspects considered are:

The purpose and character of the use, including whether such use is of a commercial nature or for nonprofit educational purposes;

The amount and substantiality of the portion used in relation to the copyrighted work as a whole; and

The effect of the use upon the potential market for or value of the copyrighted work.

The plaintiffs in the various lawsuits are a mix of creatives, such as comedians and artists, and corporate rights-holders, such as the New York Times and Getty Images. While each complaint is unique, all the cases assert that the use of copyrighted works without permission or payment is a willful infringement of copyright law, and that model developers must pay significant fines, destroy any models that were trained on such materials, and commit to not using copyrighted data in training in the future.

It is worth noting that this response is not new in some respects. Responses to new technologies are often severe, especially when incumbents and politically connected groups face disruptive or uncertain futures. In the early 2000s, file-sharing platform Napster was sued by record labels for its role in the distribution of pirated music. This litigation ultimately bankrupted the company and halted the use of file-sharing broadly. Not content with disrupting the use of the technology alone, the Recording Industry Association of America began suing individuals who shared or downloaded music outside official channels. Distortion, fear mongering, and attempts to halt innovation to preserve their place in the market have been hallmarks of such incumbents, making a clear-eyed and nuanced analysis all the more critical in the case of AI training and copyright.

There are strong arguments for protecting under fair use the use of copyrighted material in training, and a line of jurisprudence stretching back more than 50 years can be instructive on the questions being raised today.

Three key themes from successful fair use defenses apply to AI training:

Fair use allows people to share information, helping to uncover new insights and uses.

Fair use protects the right to learn and iterate upon prevailing ideas.

Fair use ensures a counterweight to the monopoly on ideas that copyright creates.

The common threads among these cases are the expansion of access to information, an opportunity to create new products or services, and the freedom to learn and build upon existing knowledge to create anew. When considering the promises of AI diffusion and improvement and how the law should treat this budding technology, these ideas should be front of mind.

Current copyright lawsuits, however, threaten such promises and risk stunting AI progress in the United States and boosting the standing and competitiveness of the Chinese AI apparatus.

4. Costs if U.S. Falls Behind in AI Development

Copyright infringement fines can range from $750 to $30,000 per infringement, with maximum penalties of $150,000 for willful infringement. Given the size of AI training datasets, successful lawsuits could result in fines totaling billions or even trillions of dollars, potentially bankrupting leading AI model developers. Beyond the monetary fines, some plaintiffs are asking courts to require the re-training or destruction of AI models whose training data included copyrighted works.

The second-order effects of adverse judgments would impact the entire AI stack, including:

: Reduced demand for high-performance semiconductors and custom computing hardware, undermining U.S. investments in semiconductor production.

: Diminished investment in data center infrastructure, which currently exceeds $400 billion annually.

: Stifled innovation in AI applications and startups. In the private sector, leading startup accelerator Y Combinator has reported that more than seventy-five percent of its summer cohort were focusing on AI models and tools, and fifty percent of its most recent cohort relied on AI models and tools to support their operations.

4.3 Military and National Security Implications

The Department of Defense (DOD) already procures systems from commercial firms or open-source models and fine-tunes them for military capabilities. Specifically, models from Meta and OpenAI are being integrated and deployed to support American military and intelligence operations. Legal restrictions on AI development would inhibit the government from accessing and deploying the private sector's most advanced models, putting American military and intelligence communities at a disadvantage compared to Chinese counterparts.

4.4 Opportunity Costs and Free Expression Impact

Beyond direct investment losses, there are significant opportunity costs in fields poised for AI-driven efficiency gains, such as drug development, physical infrastructure, and transportation. Copyright maximalism could also restrict the use of AI tools for individual expression, criticism, and speech.

5. How Copyright Restrictions Benefit China

Chinese model developers operate under two sets of rules: strict regulations for "public use" within China, and more permissive conditions for research, enterprise, or industrial uses. This dual approach allows China to control domestic information while enabling Chinese firms to compete globally.

China has exempted model developers from data and content restrictions, including copyright limitations, if the model is designed for "enterprise, education and research, or industry association." This gives Chinese developers significant advantages in leveraging vast amounts of data and an incentive to partner with industry and research organizations to protect investment and scale the technology.

Chinese model development is and will continue to support the CCP’s strategic and military might. The CCP's expansive surveillance capabilities are being enhanced with AI to monitor dissent. Chinese firms are producing innovative products combining AI models with data to improve sensors, radar systems, and computer vision—all dual-use technologies with both civilian and military applications.

If the United States loses its leadership position in AI, China will fill the void, with negative consequences domestically and globally. Unlike in China, American AI developers do not have to abide by government-imposed speech codes or socialist values, and they can choose whether to partner with the state. Allowing Chinese AI models to be developed and diffuse more quickly will harm both American soft and hard-power.

The legal battles around AI and copyright raise the specter that a domestic, commercial squabble of interests might prevent the United States from retaining its edge in the global competition around the technology. The victory of the litigants in the many pending lawsuits against AI companies may lead to legal regimes that ultimately impose ruinous monetary damages, allow the extraction of burdensome settlements, and in extreme cases result in the court-ordered destruction of AI models themselves. All would do great harm to the emerging market of products and services leveraging the technology.

Regardless of the outcome, the persistent uncertainty around the legal standing of training raises real costs, as well. In all likelihood, pending lawsuits that appear likely to set precedent in this area will take years to resolve. Facing this uncertainty, leading AI developers may very well choose to rebase their model development operations in countries with more favorable environments for training. Other nations, such as Japan and Singapore, have explicitly carved out exceptions for the training of AI systems, and others could follow suit in an effort to capitalize on a sustained period of American inaction.

The state of the research field and the market for AI services is evolving fast enough that the United States cannot afford to wait on the slow process of judicial policymaking in this area. Properly constituted, copyright can serve not as a threat to technological progress and American advantage, but as an asset. I propose four immediate policy actions to this end.

6.1 Congressional Clarification of Fair Use

Congress has the power to settle the copyright status of training immediately. It should pass a bill that asserts unequivocally that all forms of training AI systems are a fair use, providing these activities with a clear mandate to move forward. Such a definitive approach would:

Immediately lower the risk of engaging in model training, encouraging new entrants to join the market

Counter the existential risk posed by pending lawsuits against leading American AI companies

Allow American AI companies to continue growing toward global dominance

6.2 Secure Training Data for American Industry

Training data—particularly high-quality training data—is an increasingly scarce commodity even as it remains critical for advancing AI capabilities. Global competitiveness depends not just on access to talent and advanced semiconductors, but also on securing the best datasets.

The U.S. Trade Representative should prioritize bringing the full weight of American commercial power to bear on opening up and securing valuable sources of training data worldwide for U.S. firms. This may include:

Trade deals that carve out fair use-style exceptions for model training in the intellectual property frameworks of other nations

Agreements to obtain exclusive access to valuable datasets in other countries, including multilingual materials, scientific corpora, and multimodal datasets

6.3 Establish the U.S. Bureau of Data Management

Some of the most valuable untapped datasets for AI advancement exist within the government. The formation of a U.S. Bureau of Data Management (BDM) within the Department of Commerce could accelerate U.S. leadership in AI.

The BDM would serve as a dedicated team of regulatory experts and technologists tasked with:

Identifying and "liberating" valuable government datasets that could be open-sourced or licensed to AI developers

Serving as a clearinghouse for archival collections, legal codes, weather data, and other valuable information

Gathering feedback from industry about the most valuable datasets to prioritize

Working with state and local governments to make their data available for AI training

Copyright is being used as a lever for parties seeking to defend the right of individuals to "opt-out" of having their creative work used in AI training. While this is an important right to preserve, litigation is not the best method to achieve this end.

Clear technical protocols and strong norms can help foster an ecosystem where opting out is easily indicated and respected. "Robots.txt," an informal voluntary standard that indicates whether a website can be indexed by bots, demonstrates this can work at scale.

The government should establish a task force bringing together technology companies, academic researchers, rights-holders, and technical experts to establish and promote a common standard allowing creators to clearly indicate when copying for AI training is permitted or not.

Debates around copyright and intellectual property have existed since America's founding, evolving to balance incentives for creation with access to information to promote progress. AI's emergence has added a new dimension to these questions, as it represents a general-purpose technology with the potential to reshape economies, militaries, and societies.

The United States is positioned to lead in AI development and deployment, but China is actively working to surpass American capabilities. Copyright lawsuits currently present the greatest impediment to American AI progress. To ensure continued U.S. leadership, clarity around accessing data and training models under the fair use doctrine is paramount.

The policy recommendations outlined above would provide immediate relief to the AI industry while securing America's competitive advantage in this critical technology domain. By taking these steps, the United States can maintain its technological edge, strengthen national security, and promote economic prosperity in the AI era.

Comment on Request for Information on the Development of an Artificial Intelligence (AI) Action Plan

Today, I submitted a comment in response to the National Science Foundation and Networking and Information Technology Research and Development Program’s request for information regarding the development of an Artificial Intelligence Action Plan. 

Click here to download a pdf of the full comment

Thank you for the opportunity to comment on the NSF and NITRD NCO’s request for information (RFI) on the Development of an Artificial Intelligence Action Plan.

The United States is the world leader in Artificial Intelligence (AI), and not by accident. America’s dominance in AI is an outgrowth of the strength and dynamism of our private sector, particularly as it relates to software, the internet, and other information technologies. The strength of our technology sector is, in turn, the byproduct of farsighted legal frameworks that encouraged both the commercialization of the early internet and the massive investments in communications infrastructure that the digital economy needed to scale.

As with the internet, securing America’s dominance in AI will require accelerating the development and diffusion of AI technologies, unencumbered by unnecessary regulation, alongside an aggressive build-out of the energy and data center infrastructure needed to train and deploy advanced AI systems at scale. As the defining technology of our era, becoming “second place” in AI is simply not an option. AI has the potential to positively transform virtually every aspect of industry and society, drive unprecedented rates of productivity growth and scientific discovery, and provide the leading nations in AI with decisive geopolitical advantages. Enhancing America’s dominance in AI is thus not merely important for economic competitiveness, but critical for U.S. national security and, ultimately, the preservation of the American way of life.

However, the AI revolution is different from the internet in several crucial respects:

 and most obviously, the U.S. became the dominant player in information technology and the internet during a period of relative unipolarity. That first-mover advantage allowed America to establish rules of the road for internet governance that both advanced U.S. interests and exported core American values, such as openness, transparency and free expression. Today, in contrast, progress in AI is occurring against the backdrop of an intense technological competition with China

 – a communist regime that oppresses its own people, bullies U.S. allies, and conducts all manner of economic and cyber warfare against American companies and citizens.

, while the internet revolution was primarily won in the lightly-regulated realms of software and microelectronics, winning the AI race will require overcoming bottlenecks in the heavily-regulated physical world, from the energy generation needed to fuel data center growth, to AI’s countless applications in robotics, material science, biology and beyond. Indeed, as a side-effect of deindustrialization, the last 40 years have seen the United States deepen its comparative advantage in high value-added “knowledge” sectors, such as software development, entertainment, higher education, management, law and finance – the very sectors AI is most likely to disrupt and deflate in the coming years. Thus, without concerted efforts to reindustrialize and deregulate the physical economy – energy, transportation, manufacturing, construction, health care, etc. – America’s “innovation in bits” risks becoming China’s “innovation in atoms.”

, progress in AI is rapidly accelerating towards systems that match or surpass expert human performance in arbitrary domains; what is often referred to as Artificial General Intelligence (AGI). Weaker forms of AGI (i.e. systems that could substantially supplant cognitive labor performed behind a computer) are now forecast to emerge as early as 2026,

 while “strong” AGI (i.e. robustly superhuman robotic systems capable of, among other things, assembling a car from scratch using only human readable instructions) is forecasted for 2030.

 Given continuously falling computational costs, the proliferation of powerful AI capabilities and highly intelligent autonomous systems is likely unavoidable over time, presenting sui generis risks to public safety and national security. While expert opinion on the imminence and severity of these risks varies widely, a large dose of humility is no doubt warranted. We are entering truly uncharted territory, and should thus adopt policy stances rooted in realism and evidence that are robust to a wide range of possible scenarios.

These three considerations – geopolitical competition with China; barriers to adoption in the real economy; the sui generis nature of AGI – point to the importance and urgency of the AI Action Plan, and serve to frame the policy recommendations that follow.

America’s current lead in AI is a function of having the world’s leading AI companies, chip designers and hyperscale compute providers. As the AI build-out continues, maintaining this lead will require concerted policy actions to address supply-chain bottlenecks in power generation and associated infrastructure. Consider that China added a record 429 GW of net new energy generation to their grid in 2024, a 21% year-on-year increase.

 In contrast, total electricity generation in the U.S. has been essentially stagnant since 2008, with new capacity from renewable energy sources offset by reduction in non-renewable energy production. Addressing the regulatory barriers to U.S. compute and energy dominance is essential to U.S. dominance in AI over the medium term, and of sufficient urgency to justify the use of national security exemptions and emergency powers.

Prioritize the rapid deployment of new baseload energy sources for AI scaling.

Replace Executive Order 14141, “Advancing United States Leadership in Artificial Intelligence Infrastructure,” with an analogous Executive Order that a) strikes the clean energy and carbon capture sequestration requirements, and b) establishes “Special Compute Zones” (SCZs) for expediting the build-out of 5GW+ AI clusters on federal lands.

Direct CEQ and DOE to implement National Environmental Policy Act (NEPA) Categorical Exclusions for AI data centers, co-located energy sources, and associated applications (e.g. small modular nuclear reactors; enhanced geothermal; power plant modifications).

Invoke Title I DPA authorities to accelerate production of critical data center components; and Title III authorities to create public-private lending arrangements for deploying new energy generation.

Develop Programmatic Environmental Assessments (PEAs) and Programmatic Environmental Impact Statements (PEISs) for new data center construction while redirecting the internal legal capacity at the DOE’s Renewable Energy Coordination Offices (RECOs) toward AI infrastructure deployments.

Implement national security exemptions across major environmental statutes to expedite deployment of energy and data center infrastructure, including the Endangered Species Act (16 U.S.C. § 1536(j)), Clean Air Act (42 U.S.C. § 7418(a)), Clean Water Act (33 U.S.C. § 1323(a)), and NEPA (40 C.F.R. § 1506.11).

Use Presidential National Emergency powers to expedite the construction of secure AI data center(s) for use by the Department of Defense under 10 U.S.C. § 2808 and P.L. 108-136, Section 318(a).

Appoint a “Data Center Czar” to advise and coordinate the above initiatives and to liaison with private sector stakeholders.

The current US export control regime has failed to curtail China’s access to frontier AI capabilities. GPU sales to China and the broader Sinosphere – Singapore, Taiwan, Malaysia – now accounts for just under half of Nvidia’s revenue by geography, roughly matching aggregate sales to the United States. Some fraction of export restricted chips sold in this region are directly resold into China through shell companies and smuggling rings. The remainder go to data centers in third countries that Chinese firms can access remotely through the cloud. These enforcement gaps will only grow in significance overtime with the rapidly growing volume of chips produced above the controlled performance threshold – precisely when the controls ought to begin binding.

Despite these gaps in enforcement, export controls have tangibly limited the build-out of China’s domestic AI compute capacity, giving the United States a structural advantage in both AI training and inference at the ecosystem level. Indeed, had export controls had not been imposed on China in 2022 and 2023, they may very well have already taken the lead in AI given their weaker energy bottlenecks. With China announcing a $1 trillion yuan ($138 billion USD) state-backed data center fund, strengthening the Bureau of Industry and Security’s enforcement capacity has thus never been more urgent.

Prioritize BIS modernization in the President’s Budget. BIS operates with outdated technology, but could fulfill their modernization plan with additional appropriations of roughly $25 million annually for five years. IT modernization is needed to enhance BIS’s analytical and enforcement capabilities, but would also benefit US exporters by accelerating the license application process. Further explore data and intelligence sharing arrangements between BIS and U.S. intelligence agencies.

Expand both BIS’s in-house technical talent and regional inspection capacity. As it stands, BIS has only two in-house Mandarin speakers, two regional inspection officers for all of China, and only one inspection officer for the rest of Southeast Asia. Regional units could be further supplemented by secondees from partner countries.

Add Nvidia’s H20 and B20 to the list of controlled chips to account for the rise of reasoning models like DeepSeek R1, which make greater use of inference compute.

Create a certification program for chip distributors that mandates the use of pre-approved logistics providers committed to notifying BIS investigators of suspicious entities.

Launch a chip smuggling whistleblower program with awards pegged to a percentage of the resulting fines.

Order a feasibility study for delay-based location verification and the creation of a centralized BIS registry for exported chips. Delay-based location verification is non-trivial to spoof and would greatly enhance BIS’s capacity to verify end-users, identify smuggling intermediaries, and geolocate chips in restricted countries. Location verification and related mechanisms for enhanced end-user controls could be piloted under the Diffusion Rule as a condition for exporting chips to countries suspected of facilitating chip diversion, such as Singapore and Malaysia. No such mechanisms should be required for chips sold to U.S. end-users.

Restrict China’s ability to access export controlled chips remotely through cloud service providers in Tier 2 countries by deeming them illegal exports in-kind (as proposed in Rep. Michael Lawler’s Remote Access Security Act).

Direct the State Department to develop a formal “AI Talent Network” among Tier 1 countries (or a new, G20-style organization) to attract AI talent to the United States and promote multilateral engagement on AI supply-chain vulnerabilities.

The capabilities of frontier AI systems are improving at an accelerating rate. Five years ago, Large Language Models (LLMs) were barely able to generate coherent English text. Today, they can program entire games and applications from a single prompt, write in-depth research reports, and reason through high-level math problems. Breakthroughs in LLM post-training through reinforcement learning have further revealed new, potentially unbounded dimensions for model scaling. Post-training will scale rapidly over the coming year, and is anticipated to make AI systems significantly more agential and superhuman in any domain where success can be objectively benchmarked or verified.

With AGI in sight, monitoring the status of frontier AI capabilities in real-time should be a national security imperative of the U.S. government. Consider that OpenAI and Anthropic both recently indicated that their best models are on the cusp of acquiring “dual-use,” Chemical, Biological, Radiological, and Nuclear (CBRN) capabilities that, according to their own policies, will need to be withheld from public deployment without robust model safeguards. Developments at the frontier can thus provide policymakers and national security advisors with foresight into the sorts of capabilities that will eventually proliferate through open weight models as the cost of computation falls overtime.

In contemporary policy discussions, AI safety and security are unhelpfully conflated with “AI ethics,” which often reduces to making AI models parrot politically correct viewpoints. Yet powerful AI systems present many bona fide, near-term risks to U.S. national security and the health and safety of ordinary American citizens. The AI Action Plan should take care to delineate policies aimed at managing the potential catastrophic risks presented by powerful, highly autonomous AI systems from more prosaic or evergreen issues, such as bias and discrimination.

Support the U.S. AI Safety Institute (AISI) in its mission to develop voluntary AI standards and evaluations for national security-relevant risks. Explicitly require AISI to evaluate open weight foreign models, such as DeepSeek, in addition to American-made models.

Adopt a policy of “Open Source Diplomacy” by giving AISI an explicit export promotion mandate. AISI can attest to the security and reliability of American open weight models for use in critical infrastructure by independently auditing training code and datasets for backdoors and data contamination. The AISI consortium’s 280 members could then be reimagined as de facto embassies for AI systems underwritten by American standards, serving as a bulwark against both the Brussels Effect and China’s efforts to export digital authoritarianism.

Designate a senior member of the National Security Council staff to coordinate the creation of an AI Operations Center – a “Situational Awareness” room – to monitor frontier AI developments and develop real time risk assessments in collaboration with DOD, DHS, FBI and ODNI. Employ the Defense Production Act to obtain up-to-date information from industry leaders under Title VII, Section 705. The AI Operations Center should serve as the 24 hour counterpart to the AISI’s Testing Risks of AI for National Security Taskforce (TRAIN).

Direct AISI and the National Labs to launch an industry collaboration to develop rigorous benchmarks operationalizing “dual-use” thresholds for CBRN capabilities. Publish a real-time tracker of the current state-of-the-art (SOTA) for both closed and open weight models along each capability dimension and by model size.

Reinstate a modified version of Executive Order 14110 § 4.2., “Ensuring Safe and Reliable AI,” to require information disclosures to the Federal Government from companies “developing or demonstrating an intent to develop potential dual-use foundation models.” Replace the floating-point operations (FLOP)-based thresholds used to define covered models with a capability-based metric derived from the above benchmarks, e.g. only require disclosures for models that significantly advance the SOTA.

AGI-like systems and deep learning techniques more generally have the potential to massively accelerate scientific discovery and technological R&D. The U.S. Federal Government can enable AI-assisted science and R&D by expanding access to its many science-relevant datasets; by promoting the values of open science in federally-funded research; by making strategic investments in core research infrastructure; and by reforming the scientific grantmaking process itself.

Consider that the Department of Energy has already proposed creating AI foundation models for materials science as part of its Frontiers in Artificial Intelligence for Science, Security and Technology (FASST) initiative. Unfortunately, outdated IT infrastructure and data access and retention policies make sharing and combining different datasets challenging under the status quo. Assembling large DOE physics and chemistry datasets into formats and within technical infrastructure suited to training foundation models has been estimated to cost around $100 million, but could potentially be achieved for much less via private sector partnerships. The DOE’s new Foundation for Energy Security and Innovation (FESI) should prioritize establishing fellowships and public-private partnerships to support such initiatives, both financially and with human capital.

At an estimated cost of $20-40 million, the DOE should further pursue the construction of robotic labs (also known as “self-driving labs”) that automate the manual process of performing experiments, allowing scientists to synthesize, validate, and characterize new materials 24 hours a day at dramatically lower costs. Such labs could be used internally by the DOE and/or rented to research institutions to ensure public datasets are used to support American researchers and promote open science. At a minimum, self-driving labs should be added as an official priority of the DOE’s preexisting FASST initiative.

AI also has the potential to transform the scientific enterprise more broadly. AI models fine-tuned for peer review are already close to parity with human reviewers. The NSF and NIH should aggressively pilot the use of AI to accelerate the peer review process, identify and reduce publication bias, and improve the quality of grant making more generally. This can be paired with reforms to grant policies to encourage open data.

Zealously defend fair use protections for AI models trained on copyrighted data as a matter of national security.

Require all publications stemming from federal grants to be deposited in open repositories at the time of publication (or even preprint stage) by default, rather than allowing any length of embargo. Further enforce Creative Commons (CC-BY) or public-domain dedications (CC0) for datasets produced with federal funding.

Establish or expand national data repositories (like NIH’s data repositories, but broader and more specialized for AI use cases), including through the proactive acquisition and storage of large scientific datasets under open licenses that explicitly allow reuse in AI training.

Make data-sharing plans a strict requirement in all grant proposals, including details about how the data will be prepared for open-access reuse (metadata format, curation standards, etc.). Further require that all federally funded datasets be registered in an open, centralized index for automatic discoverability by AI researchers.

Revise Intellectual Property and data licensing rules. For example, consider waiving or modifying certain provisions of the Bayh-Dole Act that give universities ownership of federally funded IP, ensuring that key research artifacts (data, code) default to open licenses. Such licenses should explicitly allow for commercial use to foster broad innovation in both academia and industry.

Explore reforms to “unbundle the university” from the research enterprise. An OSTP rulemaking could relax institutional eligibility requirements mandated under the NSF’s “Proposal & Award Policies & Procedures Guide” and NIH’s “Grants Policy Statement.” Agencies could then establish new programs to allow qualified individual scientists or ad-hoc consortia to apply for federal grants directly, bypassing the standard requirement for an institutional sponsor. OSTP could similarly issue memos directing agencies to experiment with alternative grant mechanisms, such as Focused Research Organizations or research collectives.

OSTP should systematically review the compliance burdens afflicting the traditional research enterprise, including as they relate to accounting rules, grant formats, Institutional Review Boards (IRB), and IP obligations, and advise the President on issuing updated research guidelines to relevant subagencies (e.g. the Office for Human Research Protections in the case of IRB enforcement).

Institutional destabilization and displacement are perhaps the most under-rated sources of near-term AI risk. The Arab Spring was potentiated by the internet and social media, for example, demonstrating that mere information technology can facilitate a regime change under certain conditions. A world with billions of superintelligent, highly autonomous AI agents presents an analogous set of risks – not necessarily through misuse, but through appropriate use at unprecedented scale.

Democratized access to AI lawyers could overwhelm the court system; AI-enabled drug design and discovery could overwhelm the FDA clinical trial process; regulatory comments written and submitted by AIs could overwhelm agency rule-makings. Across the board, a world with democratized AGIs promises to render many legitimate functions of government inoperable due to “Denial of Service”-style attacks, like an email inbox without a spam filter. Just as AI voice cloning forced financial institutions to adopt new policies around voice verification, Federal agencies must systematically review their institutional readiness for a world with AGI.

This class of risks is manageable if and only if the rate of AI diffusion within the public sector keeps pace with the private sector. In contrast, the Biden Administration’s OMB Directive on AI established a multi-part approval process for deploying even the most rudimentary AI systems within government. The Trump Administration should take the opposite approach by empowering the Department of Government Efficiency (DOGE) to drive AI adoption within the Federal government in collaboration with OSTP, OMB and agency CIOs. For example, attorneys at the FTC could be empowered to use open source and securely hosted AI models in the discovery process. While many agencies will have bespoke AI needs or the need for fundamental process reforms to unlock AI use cases, the promise of AGI lies in its generality. A small number of generalist models should thus be authorized for use government-wide, analogous to how FedRAMP consolidated authorizations for cloud services.

At the same time, the potential of AI for monitoring and enforcement risks creating a world of “perfect compliance,” surveillance and censorship. Consider the NHTSA’s recall of Tesla FSD for performing rolling stops – something drivers do every day, and which is primarily illegal so that traffic enforcers can police edge-cases. Care must therefore be taken to ensure government use of AI remains privacy and civil liberties preserving, including from the incidental increase in the stringency of existing law and regulation through exponentially improved enforcement.

Direct the General Services Administration to restore a version of FedRAMP’s Emerging Technology Prioritization program as an interim measure to encourage the adoption of generative AI in government.

Streamline the FedRAMP authorization path for AI-specific services or updates that are low-risk or highly targeted (e.g., pilot uses of language models), and allow “sandboxing” or “pilot” designations within FedRAMP for AI tools with narrow compliance scopes.

Strengthen FedRAMP’s “Authority to Operate” (ATO) reciprocity clause so that once an AI cloud service is approved at a certain impact level, other agencies can leverage that authorization without re-opening the accreditation process.

Review FedRAMP and broader IT procurement guidelines for opportunities to promote adoption of open source AI and reduce compliance burdens that benefit Big Tech incumbents.

Leverage the Technology Modernization Fund to rapidly modernize critical IT systems for AGI readiness.

Systematically review the potential for AI-enabled mass surveillance through existing regulatory vectors, and develop recommendations for safeguarding against overreach.

AI systems now seem likely to surpass human performance across most cognitive domains within the next year or two. By its very nature, Artificial General Intelligence has the potential to directly substitute for human intelligence and agency, including in domains that require forms of creativity, strategic judgement, and emotional intelligence that are ostensibly unique to humans. Moreover, the extent to which AI augments or substitutes for human labor will be largely determined by competitive market dynamics largely beyond the control of public policy. Maintaining “a pro-worker growth path for AI,” in the words of Vice President JD Vance, is thus a worthy aspiration but far from a fait accompli.

Technological innovation has historically served to augment human labor, drive productivity growth, and ultimately create more jobs than it destroys. Yet these are long-run dynamics that can often mask severe, short-run adjustment costs. The trade shock and wave of offshoring associated with the ascension of China to the World Trade Organization comes to mind. The China Shock eliminated up to 2 million American manufacturing jobs concentrated in the rural Midwest and South Atlantic. Affected workers tended to be older and with skills specific to their industry. The new jobs that were created, meanwhile, were often in entirely different parts of the country. Dislocated workers were thus forced to turn to early retirement, disability insurance, or lower wage service sector jobs. This experience is relevant to the extent the economics of AGI mirror a sudden labor supply and offshoring shock that depresses labor demand in exposed sectors.

However, the economics of AGI differ from the China Shock in at least two crucial respects. First, AGI is likely to impact a wide cross section of jobs in various geographies and across the income scale. This is salutary to the extent that the jobs created in the wake of AI-driven automation are horizontally distributed, reducing geographical frictions to labor reallocation. And second, AI systems can be operated using natural language, and will thus likely drive education innovations that make reskilling substantially easier than in the past.

The Secretary of Labor should issue guidance clarifying that individuals whose primary occupations are at high risk of automation or AI displacement qualify for Dislocated Worker services under Title I of WIOA. Workforce Development Boards (WDBs) should be further instructed to incorporate data on AI‐related displacement risks into their local plans as required by 20 CFR § 679.560.

The Employment and Training Administration (ETA) should initiate notice‐and‐comment to update WIOA performance accountability provisions under 20 CFR Part 677 toward developing standardized metrics for tracking reemployment outcomes in AI‐impacted industries (see: 29 U.S.C. 3141). Local WDBs should be instructed to prioritize short‐term AI upskilling programs, including micro‐credentials and AI literacy training, as allowable under 29 U.S.C. 3174 and state-level "rapid response" programs.

The Secretary of Labor has substantial discretionary authority over demonstration grants and pilot programs (29 U.S.C. 3224). These authorities should be used to launch an AI Reskilling Demonstration Grant program for jumpstarting AI-assisted forms of accelerated education / retraining in collaboration with workforce boards, community colleges and private technology companies. Grants can be finely-tuned using priority scoring to ensure existing grants and E&T programs favor AI-robust skills and occupations, such as trades apprenticeships.

Given the outsized role of states in labor policy implementation, DOL should make liberal use of WIOA waiver requests under 29 U.S.C. 3249(i) to provide states with flexibility to pursue new and experimental AI‐focused workforce strategies.

Explore regulatory changes to enhance WARN Act data quality and frequency and improve reporting compliance (e.g. the creation of a standardized digital form). The WARN Act covers private‐sector employers with 100 or more employees, requiring 60 days’ notice for “Plant Closings that affect 50 or more employees within a 30‐day period, or Mass Layoffs involving at least 500 employees (or 50 employees if they constitute at least one‐third of the workforce).” This is important in part because notice of layoff under WARN is often prerequisite to qualifying as a dislocated worker under Title I.

Developing an AI Action Plan

The United States currently boasts many of the world’s leading AI model developers, ranging from established tech giants to emerging startups. But the US risks losing its competitive edge due to its own mistaken policies, which have left the door open for a slew of copyright disputes. With at least 

 active lawsuits targeting AI developers for using copyrighted materials in model training, America’s leadership in AI is under threat.

As with any new technology, disruption can manifest in new social norms, labor market churn, and institutional instability. But this should not dissuade Americans from developing and leveraging a new technology such as AI, particularly because of the potential benefits it could unlock. Copyright law is a form of industrial policy, balancing ownership to incentivize the creation of new creative works while ensuring information is available to the public to spur new innovations and uses. This is the primary reason intellectual property and copyright were enshrined in the US Constitution. To take a step toward that uncertain future, Congress should codify the use of publicly available data for AI training as fair use and kickstart the development of opt-out standards for content creators.

AI development relies on vast datasets to train models capable of diverse applications. Current legal disputes question whether the presence of copyrighted material in training data constitutes infringement. Plaintiffs are asking courts for 

 equal to the GDP of some small countries, the scrubbing of public data sets, and even the destruction of AI models. If successful, these lawsuits would create a chilling effect on the US AI industry, rendering many firms insolvent or close to it.

To Preserve US AI Leadership, Congress Must Address Copyright

Cloud computing is at the center of digital commerce. Cloud services enable firms to migrate data and access it through “the cloud,” making information available anywhere and anytime. This 

 both startups and incumbents to reduce upfront costs and access bespoke services in addition to on-demand compute, a key input for AI model development and deployment. The market for cloud services is growing steadily, with three firms at the forefront: Amazon, Google and Microsoft. These firms are colloquially referred to as 

, a term that denotes their presence across the layers of technology and infrastructure that support cloud computing operations.

But cloud computing is under attack. Regulators in Europe, the U.K. and the U.S. have all opened investigations into such firms, insinuating that antitrust action or new regulation may be warranted. While competition enforcers play an important consumer protection role in various industries, including cloud computing, imposing new regulations that would limit cloud computing providers’ ability to grow and innovate is the wrong approach. A regulatory posture that encourages investment, allows for differentiation and customization and keeps costs low would further promote competition and choice in this vital market.

Before the advent of the cloud, if a firm wanted to store data, it had to purchase servers and manage its own information infrastructure. While this gave the firm complete control over its data, it required renting space in a data center or building infrastructure internally, a significant capital investment. By contrast, cloud computing services can provide that infrastructure for firms, along with other products and services, which lowers those firms’ operating costs.

How To Promote Greater Competition for Cloud Computing

This paper examines how artificial general intelligence (AGI) could fundamentally reshape the delicate balance between state capacity and individual liberty that sustains free societies. Building on Acemoglu and Robinson's 'narrow corridor' framework, we argue that AGI poses distinct risks of pushing societies toward either a 'despotic Leviathan' through enhanced state surveillance and control, or an 'absent Leviathan' through the erosion of state legitimacy relative to AGI-empowered non-state actors. Drawing on public administration theory and recent advances in AI capabilities, we analyze how these dynamics could unfold through three key channels: the automation of discretionary decision-making within agencies, the evolution of bureaucratic structures toward system-level architectures, and the transformation of democratic feedback mechanisms.

 Our analysis reveals specific failure modes that could destabilize liberal institutions. Enhanced state capacity through AGI could enable unprecedented surveillance and control, potentially entrenching authoritarian practices. Conversely, rapid diffusion of AGI capabilities to non-state actors could undermine state legitimacy and governability. We examine how these risks manifest differently at the micro level of individual bureaucratic decisions, the meso level of organizational structure, and the macro level of democratic processes. 

To preserve the narrow corridor of liberty, we propose a governance framework emphasizing robust technical safeguards, hybrid institutional designs that maintain meaningful human oversight, and adaptive regulatory mechanisms. Our recommendations focus on practical steps to ensure AGI deployment strengthens rather than subverts democratic values, from privacyenhancing technologies to new models of citizen participation. We conclude that maintaining free societies in an age of AGI requires deliberate institutional innovation to harness its benefits while guarding against both centralized control and institutional collapse.

AGI, Governments, and Free Societies

, co-founder Yichao “Peak” Ji described the system as “the first general AI agent,” capable of doing a wide range of tasks using a computer like a human would. While numerous startups, as well as OpenAI and Anthropic, have released general computer-using agents, the company itself claims superior performance to those products. Many reports from social media seem to agree, though there are 

Manus is not available to the general public as of this writing. Monica has given access to a select group of users—seemingly focused on high-profile influencers. I was not offered an access code, but I was able to use the system for a couple of prompts. My tentative conclusion from that experience—as well as the uses I have seen from others—is that Manus is the best general-purpose computer use agent I have ever tried, though it still suffers from glitchiness, unpredictability, and other problems.

 that Manus represents another “DeepSeek moment,” where a Chinese startup is surprisingly competitive with top-tier American offerings. I suspect this analogy confuses more than it clarifies. DeepSeek is a genuine frontier AI lab. They are on a quest to build AGI in the near term, have a deep philosophical conviction about the power of deep learning, and are staffed with a team of what Anthropic Co-Founder Jack Clark has 

Why Manus Matters

CEQ Still Has Power

Big Nuke Country

On the US AI Safety Institute

The Case for Focusing FEMA’s Mission

The Permitting EOs Part 3: Shiver Me Timbers

How Should AI Liability Work? (Part II)

Job Listing: Operations Manager

Life, Liberty, and Superintelligence

Make Foreign Student Exchange Great Again 

AI, Encryption, and Data Flows: Policy Imperatives for U.S. Leadership

A Solution for Cutting Government Waste

Getting the FACTs About Foreign Adversary Investment in Telecoms

Coalition Letter in Support of Gail Slater's Nomination

The State of U.S. Science and Technology: Ensuring U.S. Global Leadership

Coalition Letter in Support of Michael Kratsios's Nomination to Lead OSTP

Testimony on House Bill 85

Coalition Letter in Support of Utah's HB 85

Technological Progress in Chinese Political Culture: An Intellectual Genealogy

On the Internet, No One Knows You're a Dog: Examining the Feasibility of Privacy-Preserving Age Verification Online

Building Baseload: Reforming Permitting for AI Energy Infrastructure

How the White House Can Reform NEPA

Copyright, AI, and Great Power Competition

Leveraging R&D and Best Practices to Advance Innovation in Education:  Recommendations for States, School Leaders, and Parents

The New Control Society

For the U.S. to remain the world leader in technology and innovation, we will need to revitalize our R&D and commercialization infrastructure—from universities to public-sector entities to venture capital funds. Securing America for the future will require managing emerging technological risks, addressing persistent challenges to critical industries, and cultivating a policy environment in which new technologies can flourish.

Technology & Innovation

As America finds itself in a new strategic competition with China, we must ensure that our technology promotes the national interest and that we protect ourselves from foreign technological threats. Answering the China challenge and countering digital authoritarianism will require the union of tech entrepreneurship, domestic policymaking, and public diplomacy.

National Security

The federal government's policy capacity has atrophied, especially in science and technology. To revive our state capacity in the legislative and executive branches and modernize how the government does business, we must strengthen the public sector's policy expertise and boost our government's capabilities and infrastructure through new technologies.

Governance

FAI is a 501(c)(3) nonprofit founded in 2014 with a mission to help bridge the gap between Silicon Valley and DC. 

Make The Future Happen Sooner.

The Policy team is a boutique think tank of policy experts and technologists focused on issues at the intersection of innovation, governance, and national security.

Explore Research Archives

Publications

Media

Fellowships

Policy Hackers

Books

Stay in the loop