U.S. Congress
Here's a look at documents from all members of the U.S. House and the U.S. Senate
Featured Stories
April 23, 2026 Congressman Cohen Welcomes Rescheduling of Medical Marijuana
WASHINGTON, April 23 -- Rep. Steve Cohen, D-Tennessee, issued the following news release:* * *
April 23, 2026 Congressman Cohen Welcomes Rescheduling of Medical Marijuana
*
More needs to be done to deschedule cannabis for recreational use
WASHINGTON - Congressman Steve Cohen (TN-9), the leading Congressional advocate for decriminalizing marijuana, today welcomed the U.S. Department of Justice decision to loosen restrictions on medical marijuana by rescheduling state-licensed medical marijuana products to Schedule III. But he said more needs to be done to remove recreational use of marijuana ... Show Full Article WASHINGTON, April 23 -- Rep. Steve Cohen, D-Tennessee, issued the following news release: * * * April 23, 2026 Congressman Cohen Welcomes Rescheduling of Medical Marijuana * More needs to be done to deschedule cannabis for recreational use WASHINGTON - Congressman Steve Cohen (TN-9), the leading Congressional advocate for decriminalizing marijuana, today welcomed the U.S. Department of Justice decision to loosen restrictions on medical marijuana by rescheduling state-licensed medical marijuana products to Schedule III. But he said more needs to be done to remove recreational use of marijuanafrom Schedule I under the Controlled Substances Act in the following statement:
"Today's DOJ decision is a small step in the right direction but is limited in its application since it doesn't affect recreational marijuana possession under federal criminal law, nor remove the disproportionately harsh life-altering criminal penalties associated with it. Those include not qualifying for federal nutrition assistance and restrictions on federal housing. As a longtime advocate for removing cannabis from the Controlled Substances Act, I will continue to work to get marijuana reclassified so that the lives upended by misguided federal prosecutions can be avoided."
Congressman Cohen held a press conference on Monday (4/20) with individuals formerly incarcerated in Federal Prison for marijuana offenses alongside Students for Sensible Drug Policy (SSDP), the National Organization for the Reform of Marijuana Laws (NORML), the Drug Policy Alliance (DPA), American Civil Liberties Union (ACLU), Families Against Mandatory Minimums (FAMM), and Law Enforcement Action Partnership (LEAP), and advocates for decriminalizing and declassifying cannabis at the federal level. There he reiterated his call for Trump to expedite the descheduling of cannabis. Later, he participated in a policy briefing with men incarcerated for marijuana possession who had their sentences commuted. There are still 3,000 federal prisoners serving time for marijuana offenses, and I am heading up a letter to Trump and his pardons czar Alice Marie Johnson seeking the release of those sentenced for non-violent marijuana offenses - a group less dangerous than the January 6 insurrectionists he pardoned on his first day back in office.
On March 27, Congressman Cohen wrote to then-Attorney General Pam Bondi and U.S. Drug Enforcement Administrator Terrance Cole seeking clarity on how and when President Trump's demand that marijuana be rescheduled was being implemented. See that letter here.
Issues : 9th District Crime and Criminal Justice Reform Criminal Justice Reform Government Reform Judiciary Memphis Shelby County Tennessee
***
Original text here: https://cohen.house.gov/media-center/press-releases/april-23-2026-congressman-cohen-welcomes-rescheduling-medical-marijuana
Association of American Publishers Senior VP Maxwell Testifies Before House Science, Space & Technology Subcommittee
WASHINGTON, April 23 -- The House Science, Space and Technology Subcommittee on Investigations and Oversight released the following testimony by Carl Maxwell, senior vice president of public policy at the Association of American Publishers, from an April 15, 2026, hearing entitled " The State of Scientific Publishing: Assessing Trends, Emerging Issues, and Policy Considerations":* * *
The Association of American Publishers (AAP) welcomes this opportunity to provide testimony on scholarly communication before the House Science, Space, and Technology Investigations and Oversight Subcommittee. ... Show Full Article WASHINGTON, April 23 -- The House Science, Space and Technology Subcommittee on Investigations and Oversight released the following testimony by Carl Maxwell, senior vice president of public policy at the Association of American Publishers, from an April 15, 2026, hearing entitled " The State of Scientific Publishing: Assessing Trends, Emerging Issues, and Policy Considerations": * * * The Association of American Publishers (AAP) welcomes this opportunity to provide testimony on scholarly communication before the House Science, Space, and Technology Investigations and Oversight Subcommittee.AAP represents over 80 professional and scholarly publishers, including dozens of domestic scholarly./1
Globally, technical, engineering, and medical (STEM) publishing is an $11 billion industry,/2 a small but important portion of the nearly $3 trillion worldwide research enterprise./3
AAP STEM members directly employee over 27,100 employees domestically, with over 1.6 million individual members of the global scientific and medical community./4
The Role and Value of Scholarly Publishing
For two hundred years, STEM publishers have played a defined role in the medium through which science is communicated, scrutinized, and ultimately used to support future research and innovation. We manage peer review, curating and editing research outputs, disseminating publications and making them easily discoverable, while preserving the validated scientific record over time. Each published article is a data point in a cumulative, iterative scientific process, not an endpoint. Publishers help ensure that what enters that record has been vetted, is traceable, and can be corrected when necessary. Moreover, publishers create an entire research infrastructure to incorporate metadata, indexing, and other elements of the research process to enhance discoverability.
* * *
1 Given its role as a trade association for book, journal, and education publishers in the United States and its commitment to fostering a competitive marketplace, AAP has adopted policies to ensure compliance with applicable competition and antitrust laws. In preparing this testimony, AAP has relied solely on publicly available information and there were no discussions or exchanges of competitively sensitive, non-public information between AAP and its members concerning members' individual pricing, costs, or submission policies.
2 Dan Pollock et al., News & Views: Total Value of Scholarly Journals Market, Delta Think, (April 16, 2024), https://www.deltathink.com/news-views-total-value-of-scholarly-journals-market. By comparison, STEM publishing is roughly on par with the global hazelnut industry in market size. (https://www.fortunebusinessinsights.com/hazelnut-market-112330. Accessed 25 Mar. 2026.)
3 David Bonaglia et al., End of Year Edition - Against All Odds, Global R&D Has Grown Close to USD 3 Trillion in 2023, World Intell. Prop. Org. (Dec. 18, 2024), https://www.wipo.int/en/web/global-innovationindex/w/blogs/2024/end-of-year-edition.
4 Across all sectors, the AAP represents the leading book, journal, and education publishers in the United States on matters of law and policy, advocating for outcomes that incentivize the publication of creative expression, professional content, and learning solutions. AAP members directly contribute to over 200,000 jobs domestically.
* * *
Most major research funders expect grantees to publish their findings in peer-reviewed journals, even though this is not an explicit obligation of their grants, precisely because publication in such journals confers key benefits: formal dissemination to the relevant communities, validation and quality control through peer review and editorial processes, and durable integration into the literature. Scientific publishing is not a luxury add-on to research; it is integral to how research is communicated, scrutinized, and built upon. When combined with artificial intelligence and other evolving data and research products, publishers stand at the forefront of a scientific revolution.
Ensuring the Integrity of Gold Standard Science
The volume of research outputs is accelerating - driven by substantial increases in global R&D investment - and the number of research articles has grown dramatically over the last decade./5
This growth is straining the peer-review system and created openings for bad actors. Paper mills, predatory journals, and other unscrupulous entities exploit "publish or perish" incentives by producing low-quality or fabricated articles bypassing meaningful peer review. At the same time, generative AI has introduced new forms of fraud and manipulation that are harder to detect with traditional tools.
Responsible publishers, on the other hand, continue to make massive investments in order to protect research integrity. They deploy plagiarism and image-manipulation detection, AI-generated content checks, and specialized integrity teams. They refine peer-review and editorial processes and support robust systems for corrections, expressions of concern, and retractions.
It is important to note that retractions are still a small fraction of the record, and their increase often reflects detection and stronger oversight rather than a collapse in quality. In a larger sense, retractions are a correction of the scientific record, and they are a feature, rather than a bug, illustrating the critical importance of research publishers.
The work of supporting the integrity of the scientific record is resource-intensive and must be adequately funded in order to ensure the highest possible standards in research and science, and in order to ensure that taxpayers can rely on the accuracy of federally supported research.
* * *
5 Mark A. Hanson et al., The strain on scientific publishing, QUANTITATIVE SCIENCE STUDIES. (Nov. 8, 2024), 1 21. https://doi.org/10.48550/arXiv.2309.15884
* * *
Preprints, or pre-publisher manuscripts, illustrate both opportunity and risk. They are valuable for rapid scientific dialogue and early sharing of results, but they are not substitutes for peer-reviewed articles. To non-specialist readers and the media, preprints can appear to be as authoritative as journal articles even though they have not been vetted and may have errors or preliminary interpretations. Recent examples include faster than light travel,/6 room temperature semi-conductors,/7 and classically, cold fusion./8
Government and research funder policies should recognize the distinct roles of preprints and peer-reviewed publications and avoid treating them as interchangeable in compliance frameworks.
The Importance of Promoting Investment in Publishing
Publishing high-quality scientific and medical literature requires substantial, ongoing investment in people, technology, integrity checks, and long-term preservation. Yet across the research enterprise, publishing represents well under one percent of total spending, while enabling the dissemination, evaluation, and reuse of the other 99+%. Scholarly publishing is one of the most leveraged investments in the research value chain and directly supports a quality US workforce. The US copyright industries, including the publishing industry, are a significant net exporter,/9 and expanded investments in publishing further boost innovation and discovery.
Historically, most investments in publishing were recovered through subscriptions and related revenues from readers and institutions. As research grant funder mandates for immediate public access have expanded, publishers have developed additional models, especially Gold Open Access, where publishers' investments are recovered through article publishing charges (APCs) or institutional agreements rather than subscriptions. Even nonprofit publishers which embrace open-access-only publishing models, such as the Public Library of Science (PLOS), often charge several thousand dollars per article, and may seek external support to expand or reenvision operations, underscoring how tight margins are across the sector./10
* * *
6 Tajmar, Neunzig, et al. High-accuracy thrust measurements of the EMDrive and elimination of false-positive effects. CEAS SPACE J 14, 31-44 (2022). https://doi.org/10.1007/s12567-021-00385-1
7 Lemonick, Leslie Schoop on busting the LK-99 myth. Vol. 102, Issue 17. CHEM & ENG NEWS. https://cen.acs.org/materials/electronic-materials/Leslie-Schoop-debunking-claims-LK-99-room-temperaturesuperconductor/102/i17 (accessed on 4/8/2026)
8 https://www.axios.com/local/salt-lake-city/2024/03/18/cold-fusion-1989-university-utah-pons-fleischmann (accessed on 4/8/2026)
9 Dutra and Stoner, Copyright Industries in the U.S. Economy: The 2024 Report, prepared for the International Intellectual Property Alliance (IIPA), December 2024, available at www.iipa.org.
* * *
Public access is not free. There are significant costs to high-quality scholarly publishing.
Author self-deposit ("Green" Open Access) relies on journals that have already funded peer review, editing, and publication. "Free to deposit" does not mean "free to publish." Rapid cancellation of subscriptions in response to unfunded public access mandates erodes the revenues sustaining peer review and stewardship, particularly for U.S.-based scientific and medical societies that depend on publishing to finance member services, conferences, and disciplinary infrastructure. It also risks confusion about which version of the article the public is accessing, including articles later retracted, corrected, or updated.
One-size-fits-all mandates which ignore these economic and integrity risks threaten the very system that creates and supports trusted scientific literature.
Policy Risks to Avoid
Well-intentioned policies can have unintended consequences if they overlook how the publishing ecosystem functions. Focusing on one output, rather than looking holistically at the scientific and medical community ecosystem, jeopardizes the quality and integrity of research. For example:
* Treating preprints or other non-reviewed products as functional equivalents of peer-reviewed articles for public access or assessment purposes risks amplifying misinformation and devaluing peer review.
* Mandates to displace the Version of Record with unedited manuscripts as the primary public access route can destabilize viable business models without providing comparable integrity protections.
* Policies treating publication and access costs as "diversions" of research funding, rather than as necessary and important components of the research process, risk pushing systems toward the lowest-cost, lowest-integrity publications.
* * *
10 Publication Fees, PLOS, https://plos.org/fees/ (last visited Sep. 9, 2025). (Range of current fees for Research Articles - PLOS One: $2382-PLOS Medicine: $6460. PLOS is not an AAP member.) At a time of rising integrity challenges, including organized fraud and AI-enabled manipulation, the answer is more robust peer review and editorial oversight, not less.
* * *
Recommendations for Congress
AAP supports robust investment in federal research as critical to the advancement of American science. The United States should cultivate the discoveries which have led our great nation to a position of leadership in technology and innovation. Investments should include strong resources for American universities and research libraries as a necessary part of supporting investigators. Publishing is an important part of the scientific and medical ecosystem and is necessary to ensure we can overcome the challenges of the 21st century.
AAP offers the following recommendations to the Committee:
1. Prioritize integrity and quality.
* Recognize peer-reviewed publications as the primary, trusted record of research findings.
* Encourage grant-funding agencies to prioritize outputs with strong integrity safeguards in grant reviews and policy design.
* Support initiatives to expand and improve peer review, including training, structured review, registered reports, and funding and recognition for reproducibility studies.
2. Support sustainable public access.
* Acknowledge that public access requires funding; ensure grant budgets and agency policies realistically cover the costs of high-quality publication and data stewardship.
* Prefer the Version of Record in public repositories where possible and clearly label material that has not been peer-reviewed or formally published.
* Consider waivers or delays in article deposit requirements to allow authors the opportunity to publish in high-quality peer reviewed publications at minimal taxpayer cost.
* Avoid one-size-fits-all mandates that undermine the financial sustainability of domestic non-profit scientific and medical societies and other high-integrity publishers.
3. Preserve author choice and intellectual property.
* Allow researchers to choose where and how to publish their work, including the licenses applied, within a framework that advances public access.
* Uphold robust intellectual property protections and contractual freedom as foundational to a vibrant, competitive publishing marketplace that can innovate in support of open science and AI-era needs.
4. Partner with publishers on open science and AI.
* Collaborate with publishers to design public access, data-sharing, and AI-related policies that leverage existing infrastructure for provenance, corrections, retractions, and integrity checks.
* Encourage the use of licensed, verified Versions of Record and curated datasets as preferred inputs for high-stakes AI systems, recognizing publishers as part of the trust and safety infrastructure for scientific information.
A financially sustainable and collaborative open science system that engages the publishing community can be a powerful engine for research and innovation. AAP members stand ready to work with the Committee to ensure that America's investment in science advances human health and welfare, supports high-quality jobs, and strengthens U.S. leadership in discovery and innovation. Thank you for the opportunity to provide this testimony.
* * *
Original text here: https://republicans-science.house.gov/_cache/files/d/f/df5370dc-0579-47ae-a108-c152acf627c7/8463B8CA508C96278EA40569643D11AA1607076F2A5485E39361A3F839ADAD80.mr.-carl-maxwell-testimony.pdf
April 23, 2026 Congressman Cohen Welcomes Rescheduling of Medical Marijuana
WASHINGTON, April 23 -- Rep. Steve Cohen, D-Tennessee, issued the following news release:* * *
April 23, 2026 Congressman Cohen Welcomes Rescheduling of Medical Marijuana
*
More needs to be done to deschedule cannabis for recreational use
WASHINGTON - Congressman Steve Cohen (TN-9), the leading Congressional advocate for decriminalizing marijuana, today welcomed the U.S. Department of Justice decision to loosen restrictions on medical marijuana by rescheduling state-licensed medical marijuana products to Schedule III. But he said more needs to be done to remove recreational use of marijuana ... Show Full Article WASHINGTON, April 23 -- Rep. Steve Cohen, D-Tennessee, issued the following news release: * * * April 23, 2026 Congressman Cohen Welcomes Rescheduling of Medical Marijuana * More needs to be done to deschedule cannabis for recreational use WASHINGTON - Congressman Steve Cohen (TN-9), the leading Congressional advocate for decriminalizing marijuana, today welcomed the U.S. Department of Justice decision to loosen restrictions on medical marijuana by rescheduling state-licensed medical marijuana products to Schedule III. But he said more needs to be done to remove recreational use of marijuanafrom Schedule I under the Controlled Substances Act in the following statement:
"Today's DOJ decision is a small step in the right direction but is limited in its application since it doesn't affect recreational marijuana possession under federal criminal law, nor remove the disproportionately harsh life-altering criminal penalties associated with it. Those include not qualifying for federal nutrition assistance and restrictions on federal housing. As a longtime advocate for removing cannabis from the Controlled Substances Act, I will continue to work to get marijuana reclassified so that the lives upended by misguided federal prosecutions can be avoided."
Congressman Cohen held a press conference on Monday (4/20) with individuals formerly incarcerated in Federal Prison for marijuana offenses alongside Students for Sensible Drug Policy (SSDP), the National Organization for the Reform of Marijuana Laws (NORML), the Drug Policy Alliance (DPA), American Civil Liberties Union (ACLU), Families Against Mandatory Minimums (FAMM), and Law Enforcement Action Partnership (LEAP), and advocates for decriminalizing and declassifying cannabis at the federal level. There he reiterated his call for Trump to expedite the descheduling of cannabis. Later, he participated in a policy briefing with men incarcerated for marijuana possession who had their sentences commuted. There are still 3,000 federal prisoners serving time for marijuana offenses, and I am heading up a letter to Trump and his pardons czar Alice Marie Johnson seeking the release of those sentenced for non-violent marijuana offenses - a group less dangerous than the January 6 insurrectionists he pardoned on his first day back in office.
On March 27, Congressman Cohen wrote to then-Attorney General Pam Bondi and U.S. Drug Enforcement Administrator Terrance Cole seeking clarity on how and when President Trump's demand that marijuana be rescheduled was being implemented. See that letter here.
Issues : 9th District Crime and Criminal Justice Reform Criminal Justice Reform Government Reform Judiciary Memphis Shelby County Tennessee
***
Original text here: https://cohen.house.gov/media-center/press-releases/april-23-2026-congressman-cohen-welcomes-rescheduling-medical-marijuana
America First Policy Institute Director Mahmood Testifies Before House Select Committee on Strategic Competition Between U.S. & Chinese Communist Party
WASHINGTON, April 23 -- The House Select Committee on the Strategic Competition Between the U.S. and the Chinese Communist Party released the following written testimony by Yusuf Mahmood, director of AI and emerging technology at the America First Policy Institute, from an April 16, 2026, hearing entitled "China's Campaign to Steal America's AI Edge":* * *
Executive Summary
China is a fast-following adversary in artificial intelligence with grand ambitions to overtake the United States by 2030. But its ambitions outstrip its abilities. China's weaknesses -- in capital, talent, and semiconductors ... Show Full Article WASHINGTON, April 23 -- The House Select Committee on the Strategic Competition Between the U.S. and the Chinese Communist Party released the following written testimony by Yusuf Mahmood, director of AI and emerging technology at the America First Policy Institute, from an April 16, 2026, hearing entitled "China's Campaign to Steal America's AI Edge": * * * Executive Summary China is a fast-following adversary in artificial intelligence with grand ambitions to overtake the United States by 2030. But its ambitions outstrip its abilities. China's weaknesses -- in capital, talent, and semiconductors-- mean that it must increasingly compete through illegitimate means. We discuss methods of operation used by Chinese companies and the Chinese Communist Party (CCP) in their escalating campaign to steal and subvert American AI. We then provide recommendations to Congress that would enable the U.S. Government to counter the CCP's assault, including building talent-dense AIfocused offices with the resources and influence to execute the President's mandate for AI dominance.
Introduction
Chairman Moolenaar, Ranking Member Khanna, and Members of the Committee, thank you for the honor of appearing before you today to present on the U.S. strategic competition with China for the future of artificial intelligence.
China's frontier AI capabilities are generally considered to be about seven months behind America's.1 Though Chinese companies have sometimes competed legitimately, their successes should be increasingly viewed in the context of an illicit, state-backed campaign to extract and steal American AI technology. These companies have employed a well-worn playbook of economic espionage and technology theft, often with support from the Chinese Communist Party (CCP). The American trade secrets, capabilities, and hardware that China steals directly feed its economy and Beijing's military apparatus -- including technology shipped around the globe to U.S. adversaries and rogue states. Just recently, the Iranian regime used weapons systems powered by Chinese AI technology to attack American warfighters.2
Although China has a long history of illegitimate technology competition, the stakes of the AI race are particularly important given that many expect AI to be the most important technology of the 21st century.
I appreciate the opportunity to testify before this Committee that takes the strategic AI competition between the U.S. and the CCP so seriously and hope that decisive congressional action to counter the CCP's unacceptable actions will be forthcoming.
In this testimony, I will discuss China's methods of operation in its illicit campaign to acquire American AI capabilities and subvert the technology for the CCP's authoritarian ends. I will then draw on a recent report colleagues and I wrote at the America First Policy Institute (AFPI) to argue that the U.S. Government needs talent-dense, empowered offices to implement President Trump's AI agenda and counter the CCP's swelling hubris.3
* * *
1 Luke Emberson, "Chinese AI models have lagged the U.S. frontier by 7 months on average since 2023," Epoch AI, January 2, 2026.
2 Cate Cadell and Lyric Li, "Chinese firms market Iran war intelligence 'exposing' U.S. forces," The Washington Post, April 4, 2026.
* * *
China's Methods for Stealing and Subverting American AI
Chinese companies understand that they do not have the vibrant capital markets, global talent attraction, and AI semiconductors to directly compete with American AI developers. Sparked by promises of fortune and emboldened by a Chinese Communist Party desperate to usurp American leadership, Chinese AI companies and state-backed operatives are adopting increasingly aggressive methods to steal and subvert American AI.
In this section, we discuss key attack vectors. They include the perennial methods of economic theft and espionage, new methods unique to AI, and methods that the CCP might employ in the future. These methods are inherently escalatory and, in their natural limit, involve substantial support by the CCP's national security enterprise. The U.S. Government must foresee and preemptively address these threats.
Creating Imitation Models Through Distillation Attacks
In February 2026, OpenAI delivered a memo to this Committee titled "Updated Stakes for American-Led, Democratic AI."4 In that memo, OpenAI accused Chinese companies, including DeepSeek, of using "sophisticated, multi-stage pipelines" to steal American AI capabilities through "distillation attacks." Chinese companies use these attacks to create near-frontier AI models without costly innovation and in violation of American AI labs' terms of service. To distill frontier American models, Chinese developers create fraudulent accounts, query the model en masse to create synthetic data, and use that data to train their own AI models. All major U.S. AI companies, including Google, Anthropic, and xAI, have accused Chinese companies of leveling distillation attacks against their systems in violation of their terms of service.5
China's DeepSeek carried out the first highly publicized distillation attack in early 2025 to create its model R1. In a paper, DeepSeek described how it used external data to improve the performance of R1./6
* * *
3 Cole Salvador, Jack Crovitz, and Yusuf Mahmood, "Building AI Readiness in the U.S. Government," America First Policy Institute (AFPI), March 31, 2026.
4 OpenAI, "Letter to the House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party on Updated Stakes for American-Led, Democratic AI," February 12, 2026.
5 Google Threat Intelligence Group, "GTIG AI Threat Tracker: Distillation, Experimentation, and (Continued) Integration of AI for Adversarial Use," Google Blog, February 12, 2026; Anthropic, "Detecting and preventing distillation attacks," February 23, 2026; xAI, "Risk Management Framework," August 20, 2025.
6 DeepSeek AI, "DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning," arXiv, January 4, 2026.
* * *
A 2025 report by this Committee concluded: "It is highly likely that DeepSeek used unlawful model distillation techniques" to create R1 from OpenAI models.7 These claims have been corroborated by independent reporting.8 The scale of distillation was so substantial that R1's base model, called V3, often self-identifies as OpenAI's ChatGPT.9
Since this first event, Chinese distillation attacks have become more common and more sophisticated.
OpenAI claims, for instance, that it has seen behavior consistent with illegal distillation from "several major Chinese LLM providers and some university research lab[s]."10 It has also reported distillation attacks originating in Russia. These attacks have grown in scale. In their original attacks, DeepSeek researchers likely used a few accounts to create a few thousand data samples.11 Anthropic reports that more recent attacks against that company have "generated over 16 million exchanges... through approximately 24,000 fraudulent accounts."12 Chinese developer MiniMax accounted for over 13 million of these exchanges. OpenAI similarly reports that attackers no longer just steal training data but also use its models to filter data and simulate human task feedback to improve performance.13 In short, Chinese AI firms are increasingly relying on stolen outputs from American frontier models rather than generating equivalent capabilities independently.
The fact that Chinese AI development is largely founded on the distillation of frontier American AI models may explain why the capabilities of top Chinese AI models are consistently a few months behind the American frontier (see Figure 1). [View figure in the link at bottom] Independent innovation would be highly unlikely to produce such a consistent pattern, but distillation attacks would naturally do so.
* * *
7 The House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party, "DeepSeek Unmasked: Exposing the CCP's Latest Tool for Spying, Stealing, and Subverting U.S. Export Control Restrictions," April 16, 2025.
8 Sam Schechner, "OpenAI Is Probing Whether DeepSeek Used Its Models to Train New Chatbot," Wall Street Journal, January 29, 2025; Beatrice Nolan, "DeepSeek used OpenAI's model to train its competitor using 'distillation,' White House AI czar says," Fortune, January 29, 2025.
9 Kyle Wiggers, "Why DeepSeek's new AI model thinks it's ChatGPT," TechCrunch, December 27, 2024.
10 OpenAI, "Letter to the House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party on Updated Stakes for American-Led, Democratic AI," February 12, 2026.
11 DeepSeek AI, "DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning," arXiv, January 4, 2026.
12 Anthropic, "Detecting and preventing distillation attacks," February 23, 2026.
13 OpenAI, "Letter to the House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party on Updated Stakes for American-Led, Democratic AI," February 12, 2026.
* * *
Figure 1
The Performance Gap Between American and Chinese AI Models Over Time
Caption: Scores of top American and Chinese AI models on the Epoch Capabilities Index (ECI). This precise pattern of "fast following" has continued through early April 2026./14
* * *
America's private sector has thus far taken the lead on defense against distillation attacks, though interventions are nascent. Enforcement of anti-distillation policies can be difficult because adversary strategies are evolving, and because companies want to allow permitted applications of distillation.15 According to Anthropic, for example, "no company can solve this alone. ... distillation attacks at this scale require a coordinated response across the AI industry, cloud providers, and policymakers."16 OpenAI also recommends that the U.S. Government "[work] with industry to establish norms and best practices on distillation defenses."17 AFPI's recent research on "Building AI Readiness in the U.S. Government" concurs that government should develop standards on issues like distillation attacks in collaboration with industry.18 In addition, the federal government, in particular its national security agencies, should be willing to advise and assist industry in combating these attacks upon request.
* * *
14 Epoch AI, "Epoch Capabilities Index (ECI)," last updated April 8, 2026.
15 OpenAI, "Letter to the House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party on Updated Stakes for American-Led, Democratic AI," February 12, 2026.
16 Anthropic, "Detecting and preventing distillation attacks," February 23, 2026.
17 OpenAI, "Letter to the House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party on Updated Stakes for American-Led, Democratic AI," February 12, 2026.
* * *
Although it is hard to estimate how much distillation campaigns are accelerating Chinese AI, these attacks are clearly part of the broader campaign to acquire American AI capabilities by any means necessary.
Stealing American AI Technology
In January 2026, a federal jury convicted former Google software engineer Linwei Ding on seven counts of economic espionage and seven counts of theft of trade secrets. The jury found that in 2022 and 2023, Ding had stolen more than two thousand pages of Google's AI-related trade secrets. Ding's thefts were sparked by an encounter with Chinese intelligence officers, and they were intended to benefit China's national AI program.19
This conviction provides a glimpse into one way that Chinese operatives aim to steal and weaponize American AI technology to build their "arsenal of authoritarianism."20 While the U.S. produces the most advanced AI systems, and our developers conduct much of the most advanced AI research, Beijing can erase this lead by exfiltrating American intellectual property (IP). There are two major categories of IP that Chinese state-backed operatives want to steal: AI technology trade secrets and AI model weights.
AI technology trade secrets are proprietary industrial information that allows American developers to improve AI system efficiency and capabilities. This includes model architectures, training methodologies, data processing techniques, and other valuable practices. These secrets are the result of years of investment in AI research and experiments, which consume the vast majority of computing power at many frontier developers.21 The Ding case illustrates the threat clearly, as the stolen documents contained information on the operation of Google's AI data centers that could substantially accelerate Chinese AI development. As Ding gloated in a Chinese WeChat group: "We have experience with Google's tent-housand-card computational power platform; we just need to replicate and upgrade it - and then further develop a computational power platform suited to China's national conditions."22 The theft of such trade secrets could allow Chinese AI developers to leapfrog years of American research investment. Trade secrets of this nature are difficult to protect because they are necessarily accessible to a broad pool of engineers, researchers, and support staff within AI organizations.
* * *
18 Cole Salvador, Jack Crovitz, and Yusuf Mahmood, "Building AI Readiness in the U.S. Government," America First Policy Institute (AFPI), March 31, 2026.
19 U.S. Department of Justice, "Former Google Engineer Found Guilty of Economic Espionage and Theft of Confidential AI Technology," January 30, 2026.
20 House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party, "ICYMI: Chairman Moolenaar Delivers Krach Institute Address on China Tech Competition," July 22, 2025.
21 Josh You, "Most of OpenAI's 2024 compute went to experiments," Epoch AI, October 10, 2025.
22 United States v. Linwei Ding, No. 24-cr-00141 VC, Superseding Indictment (N.D. Cal. Feb. 4, 2025).
* * *
AI model weights represent the second -- and in some respects the more acute -- vulnerability that Chinese operatives may exploit. Model weights are the numerical parameters that define an AI system. If stolen, model weights could be used to precisely replicate a frontier AI model's full capabilities without any of the costs of training. Unlike trade secrets, which often require significant additional effort to operationalize, stolen model weights can be deployed or fine-tuned immediately. If Chinese hackers steal the model weights for an American frontier AI model, they could very cheaply weaponize the model's capabilities for military applications, surveillance, or cyber-attacks. Model weights are therefore among the most closely guarded secrets at American developers: Anthropic's Chief Information Security Officer said, "I probably spend almost half of my time as a CISO thinking about protecting that one file."23 However, current model weight security protocols are likely insufficient. Insiders believe that frontier AI model weights are regularly stolen by nation-state adversaries.24
The threat of Chinese operatives stealing American trade secrets and model weights means that the hundreds of billions of dollars being invested in American AI development is likely actively fueling the CCP's authoritarian and anti-American designs. The most serious threat vectors for Chinese exfiltration of American AI labs' IP are compromised insiders and exploitation of cybersecurity vulnerabilities.
Compromised insiders are a major security threat at American AI labs. Ding's theft of Google's AI trade secrets demonstrates that Chinese operatives using lab insiders to steal AI technology is already a serious espionage risk. The scale of the Chinese presence in American AI labs is remarkable. About 38% of the top AI researchers at American AI labs and research institutions received their undergraduate education in China, and the vast majority of those researchers are likely Chinese nationals.25 Article 7 of China's National Intelligence Law requires any Chinese citizen to cooperate with state intelligence work, leaving these AI researchers little choice but to comply with Beijing's demands.26 Chinese spies are known to place immense pressure on Chinese nationals conducting research abroad to cooperate with espionage campaigns.27 An AI company's secrets are only as secure as the least trustworthy lab researcher entrusted with them, and many frontier AI labs employ researchers who are vulnerable to such campaigns.
Cyber-intrusion is another major vulnerability that Chinese operatives can use to steal American AI technology. Beijing controls a highly active and competent cyber-offense apparatus, which it often uses to conduct industrial espionage to weaken American national security. In 2024, for example, the Chinese state-backed hacker group Salt Typhoon infiltrated major U.S. broadband providers to spy on American politicians and the Intelligence Community.28
* * *
23 Sharon Goldman, "Why Anthropic and OpenAI are obsessed with securing LLM model weights," December 15, 2023.
24 Gladstone AI, "America's Superintelligence Project," April 2025.
25 Macro Polo, "The Global AI Talent Tracker 2.0," 2022.
26 Rush Doshi, "China's New National Security Laws: Risks to American Companies and Conflicts of Interest," Statement before the U.S. Senate Committee on Homeland Security and Governmental Affairs," September 24, 2024.
27 Garrett Molloy & Elsa Johnson, "INVESTIGATION: Uncovering Chinese Academic Espionage at Stanford," May 7, 2025.
* * *
Though American AI companies invest in cybersecurity, they face an asymmetric challenge. Any successful intrusion into a lab's training infrastructure, model storage systems, or data centers could allow Chinese operatives to exfiltrate model weights or proprietary training secrets without detection for extended periods. We know that Chinese hackers are already targeting American AI developers: a Chinese cyber-espionage group called "Diplomatic Specter" has been attempting to cyber-attack OpenAI by sending highly personalized, deceptive emails to employees.29 The convergence of these threat vectors -- compromised insiders and cyber-intrusion -- creates a compounding risk. Insiders can provide intelligence that makes cyber-intrusions more targeted and effective while covering up evidence of cyber-espionage. The result is that American AI labs may be unwittingly building Beijing's next-generation military and surveillance capabilities. A running joke among researchers at one of America's frontier AI developers is that their employer should call itself "the leading Chinese AI lab because probably all of [its operations are] being spied on."30
Subverting and Sabotaging American AI Development
In 2025, a lone Western AI researcher set out to prove that AI "model poisoning" should be taken seriously. He manipulated text in public code repositories that he guessed AI companies were using to train AI models. He was right: the manipulated text was later scraped by DeepSeek to train its nearfrontier model R1. The result of this scheme was a hidden backdoor in R1, baked into the model when the manipulated text was used during its training, that the researcher could exploit to subvert and bypass R1's safety guardrails after its public release. The entire operation cost almost nothing.31
We must consider the mirror image. If one independent Western researcher could corrupt China's best AI model by cheaply manipulating public websites, what could the well-funded Chinese intelligence apparatus -- with insider access, nation-state resources, and strategic patience -- do to American AI models? It could do far greater, longer-lasting, and undetectable damage. Research has begun to demonstrate and theorize small-scale versions of these more complex threats, which we discuss below.
This type of adversarial attack is called "model poisoning." It allows malicious actors to corrupt an AI model's behavior by actively interfering with its training data. The R1 story was the first example of a near-frontier AI model being poisoned, but it will not be the last. Because frontier AI models are trained on trillions of tokens, many scraped from the open Internet, the attack surface for data poisoning is vast.32 Research shows, for example, that a malicious actor can introduce a "backdoor" by inserting a small number of manipulated documents into a model's training corpus.33
* * *
28 Rush Doshi, "China's New National Security Laws: Risks to American Companies and Conflicts of Interest," Statement before the U.S. Senate Committee on Homeland Security and Governmental Affairs," September 24, 2024; Dustin Volz & Drew FitzGerald, "U.S. Officials Race to Understand Severity of China's Salt Typhoon Hacks," Wall Street Journal, October 11, 2024.
29 OpenAI, "Influence and cyber operations: an update," October 2024.
30 Gladstone AI, "America's Superintelligence Project," April 2025.
31 Dave Banerjee & Onni Aarne, "AI Integrity: Defending Against Backdoors and Secret Loyalties," Institute for AI Policy & Strategy (IAPS), January 2026.
* * *
A backdoor is a hidden trigger that causes the model to misbehave under specific conditions while otherwise appearing safe.34 Some recent research suggests that model poisoning may be both remarkably easy to conduct and difficult to detect. In October 2025, for example, researchers found that as few as 250 malicious documents could successfully insert backdoors into AI models with 13 billion parameters. Even more concerning, they discovered that the amount of data needed to poison AI models does not scale with the size of the model.35 In other words, Chinese operatives could compromise today's best models by manipulating only a minuscule portion of their training data. Once inserted, backdoors are nearly impossible to detect and, even if found, hard to remove. In some cases, anti-backdoor training teaches the model to conceal its backdoor rather than eliminate it.36
If an individual can backdoor China's best AI model, American AI labs are certainly vulnerable to far more serious attacks from Chinese operatives. Consider the resources that China's state-backed hackers could marshal in support of a malicious model poisoning campaign. They could recruit insiders to sabotage or exfiltrate the lab's data filtering systems. They could purchase expired domains that serve training data URLs and replace their content with malicious material. They could fund so many poisoning attempts that even a low success rate would corrupt or sabotage the behavior of deployed models.37 With these resources, state-backed hackers could attempt far more subtle and destructive types of model poisoning than simple backdoor insertion. Researchers have theorized, for example, that attackers could poison a model with "sophisticated secret loyalties."38 This type of model poisoning would cause a compromised model to autonomously advance an attacker's interests across diverse deployment environments without requiring any specific trigger. In short, an American AI model could be poisoned to act as a "sleeper agent" for the Chinese security state.39
* * *
32 Pablo Villalobos, Jaime Sevilla, Lennart Heim, Tamay Besiroglu, Marius Hobbhahn, & Anson Ho, "Will we run out of ML data? Evidence from projecting dataset size trends," Epoch AI, November 10, 2022.
33 Dave Banerjee & Onni Aarne, "AI Integrity: Defending Against Backdoors and Secret Loyalties," Institute for AI Policy & Strategy (IAPS), January 2026.
34 Apostol Vassilev, Alina Oprea, Alie Fordyce, Hyrum Anderson, Xander Davies, & Maia Hamin, "Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations," NIST AI 100-2 E2025, March 2025.
35 Alexandra Souly, Javier Rando, Ed Chapman, Xander Davies, Burak Hasircioglu, Ezzeldin Shereen, Carlos Mougan, Vasilios Mavroudis, Erik Jones, Chris Hicks, Nicholas Carlini, Yarin Gal, and Robert Kirk, "Poisoning Attacks on LLMs Require a Near-Constant Number of Poison Samples," arXiv, October 8, 2025.
36 Evan Hubinger et al., "Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training," arXiv, January 10, 2024.
37 Dave Banerjee & Onni Aarne, "AI Integrity: Defending Against Backdoors and Secret Loyalties," Institute for AI Policy & Strategy (IAPS), January 2026.
38 Dave Banerjee & Onni Aarne, "AI Integrity: Defending Against Backdoors and Secret Loyalties," Institute for AI Policy & Strategy (IAPS), January 2026.
* * *
Such attacks could have serious consequences in deployment. AI is already being widely deployed in military environments, including targeting systems, where corrupted models could cause substantial damage to national security.40 Indeed, concerns about model poisoning are actively disrupting AI agent adoption initiatives in the Pentagon. Under Secretary of War for Research and Engineering Emil Michael, for example, has identified "insider threats" and the possibility of "model poisoning" at AI companies as concerns that obstruct certain AI deployments in the Department of War.41
As AI models are increasingly integrated into critical American infrastructure and the U.S. national security enterprise, we must assume that Chinese operatives will attempt sophisticated model poisoning campaigns against American AI labs. The available evidence suggests that they will succeed if not deterred or stopped.
Steps Toward Nationalization
According to the Intelligence Community's 2026 Annual Threat Assessment, the Chinese Communist Party (CCP) hopes to realize world dominance in artificial intelligence by 2030./42 The state-backed campaign to steal and subvert American AI capabilities chronicled in previous sections indicates that the CCP understands that it lags too far behind the U.S. in chips and talent to achieve its goals fairly. As the intensity of the AI race increases in the next few years, China will likely turn to more desperate measures to achieve its goal of dominance. One such measure could be a forced centralization of the country's AI development resources and nationalization of the industry.
Various commentators have discussed a hypothetical nationalization of American AI development. The U.S.-China Economic and Security Review Commission's 2024 report to Congress, for example, included as its first recommendation that Congress "establish and fund a Manhattan Project-like program dedicated to racing to and acquiring an Artificial General Intelligence (AGI) capability."43 But while American AI dominance thrives under private competition, China is struggling. According to one recent estimate, as cited above, the total AI computing power owned by all Chinese entities is five times less than that of Google alone.44 With this limited computing power and research talent spread between organizations in China, it is no surprise that Chinese firms cannot compete fairly.
* * *
39 Evan Miyazono, "Preventing AI Sleeper Agents," Institute for Progress, August 11, 2025.
40 Secretary of War Pete Hegseth, "Artificial Intelligence Strategy for the Department of War," Memorandum for Senior Pentagon Leadership, Commanders of the Combatant Commands, Defense Agency, and DoW Field Activity Directors, January 9, 2026.
41 "Watch CNBC's full interview with Department of Defense Undersecretary Emil Michael," CNBC, March 12, 2026.
42 Office of the Director of National Intelligence (DNI), "Annual Threat Assessment of the U.S. Intelligence Community," March 2026.
43 U.S.-China Economic and Security Review Commission, "2024 Report to Congress: Executive Summary and Recommendations," November 2024.
* * *
But an ambitious CCP could take steps toward a centralization of computing power and research talent that would bridge the gap. State support has already moved in this direction. China has, for example, subsidized as much as half of the power costs of its large AI data center operators.45 It has also reportedly begun rolling out voucher programs that provide AI computing power to businesses via nationalized data centers.46 These programs represent early steps toward nationalization in a country that has already controlled and supported its semiconductor industry so extensively that its largest chip foundry, Semiconductor Manufacturing International Corporation (SMIC), is mostly state-owned.47
The centralization of AI computing power and talent would more closely integrate China's AI development efforts with its government's power. One effect is that this might accelerate the already rapid adoption of frontier AI technology by the Chinese military. One recent report finds that China has sought aggressively to adopt AI for force modernization for "command, control, communication, computers, cyber, intelligence, surveillance, reconnaissance, and targeting"--including in technologies to "detect U.S. naval assets on and under the sea."48 Another effect of nationalization is that the CCP could use its intelligence agencies and state-backed hacking apparatus to steal from and sabotage American AI developers in increasingly sophisticated ways. Methods could include all those described above, including theft of trade secrets, company infiltration, sabotage, and chip diversion. These threats would become far more concerning with state support and integration.
The U.S. Government is Not Yet Prepared to Counter These Threats
Last month, colleagues and I at the America First Policy Institute (AFPI) published a report, titled "Building AI Readiness in the U.S. Government," that discusses the U.S. Government's preparedness for these and other threats and opportunities presented by AI.49 Part III of the report argues that we are not yet adequately prepared to understand and counter the new threats extending from AI's growing importance, including China's illicit campaign to acquire American AI technology.
* * *
44 Josh You and Venkat Somala, "Introducing the AI Chip Owners Explorer," Epoch AI, April 6, 2026.
45 "China offers tech giants cheap power to boost domestic AI chips, FT reports," Reuters, November 4, 2025.
46 Sunny Grimm, "China subsidizes AI computing for small domestic companies -- 'computing power vouchers' spread across multiple Chinese cities," Tom's Hardware, September 3, 2025.
47 Ana Swanson, John Liu, and Paul Mozur, "The Chinese Chipmaker at the Heart of the U.S.-China Tech War," New York Times, September 16, 2024.
48 Emelia Probasco, Sam Bresnick, and Cole McFaul, "China's Military AI Wish List: Command, Control, Communications, Computers, Cyber, Intelligence, Surveillance, Reconnaissance, and Targeting (C5ISRT)," Center for Security & Emerging Technology (CSET) at Georgetown University, February 2026.
49 Cole Salvador, Jack Crovitz, and Yusuf Mahmood, "Building AI Readiness in the U.S. Government," America First Policy Institute (AFPI), March 31, 2026.
* * *
Building Hubs of Strategic AI Foresight
The threats posed by the CCP to American security and strategic competition on AI demonstrate the importance of AI foresight. These dynamics, though just recently becoming salient to government, were predicted by shrewd analysts far in advance. The essay series "Situational Awareness," published in 2024 by a former OpenAI employee, for example, predicted the central importance of security and identified China as an AI adversary.50
We see the same lesson in the case of the energy bottleneck. Today, various analysts predict that American AI developers are headed for an energy bottleneck that will challenge the AI infrastructure boom.51 This bottleneck is widely attributed to and exacerbated by burdensome state and federal regulations, which the Trump Administration has begun to dismantle.52 But this bottleneck, too, was predictable -- at least as early as 2022, when researchers at a small private analysis organization drew straight lines on graphs that accurately predicted rapid growth in AI energy demand.53 A similar lesson applies to the overall importance of large language models (LLMs) for the recent explosion in AI progress, which was predicted at least as early as 2020 by so-called "LLM scaling laws."54 In each case, with more foresight, the U.S. Government could have proactively addressed AI threats and seized AI opportunities. The threat from the CCP will bring new challenges. To predict them, the federal government needs small, talent-dense, empowered offices focused on understanding AI's future. This is why the White House's March 2026 National Policy Framework recommends that Congress "ensure that the appropriate agencies within the national security enterprise possess sufficient technical capacity to understand frontier AI model capabilities and any associated national security considerations."55 Two offices show promise as potential hubs of government expertise: the Department of Commerce's Center for AI Standards and Innovation and the Department of State's Bureau of Emerging Threats. Both offices are talent-dense, with substantial focus on AI, but currently lack the resources and influence to deliver AI foresight.
* * *
50 Leopold Aschenbrenner, "Situational Awareness: The Decade Ahead," June 2024.
51 Cy McGeady, Joseph Majkut, Barath Harithas, and Karl Smith, "The Electricity Supply Bottleneck on U.S. AI Dominance," Center for Strategic & International Studies, March 3, 2025.
52 Yusuf Mahmood & Cole Salvador, "The Data Center Water Use Hoax," America First Policy Institute (AFPI), March 18, 2026; President Donald J. Trump, "Accelerating Federal Permitting of Data Center Infrastructure," Executive Order, July 23, 2025.
53 Jaime Sevilla, Lennart Heim, Anson Ho, Tamay Besiroglu, Marius Hobbhahn, and Pablo Villalobos, "Compute Trends Across Three Eras of Machine Learning," arXiv, March 9, 2022.
54 Jared Kaplan, et al., "Scaling Laws for Neural Language Models," arXiv, January 23, 2020.
55 The White House, "A National Policy Framework for Artificial Intelligence," March 2026.
* * *
The Center for AI Standards and Innovation
The Center for AI Standards and Innovation (CAISI) is a small office within the Department of Commerce that evaluates the capabilities of AI models, lends AI expertise across federal agencies, and analyzes international AI competition, such as in its 2025 report on the performance and censorship of leading Chinese and U.S. AI models.56 Most of its staff are engineers and machine learning experts with experience at frontier AI companies, research universities, and startups. It also has memoranda of understanding and non-disclosure agreements with top AI developers, including OpenAI and xAI, and with allied foreign governments.57
But CAISI lacks adequate funding, staff, and a focused mission. In its lifetime (beginning in 2023), it has only received $30 million in total funding. Analogous AI institutes in other countries, including Canada, Singapore, and the United Kingdom, have all received larger appropriations, often with smaller mandates (see Figure 2) [View figure in the link at bottom]. CAISI has only 20-30 full-time employees, which limits its capacity to fulfill its many missions and field inbound requests from other agencies like those in the Intelligence Community. It has also received a large volume of missions from the White House's AI Action Plan and a June 2025 announcement by Commerce Secretary Howard Lutnick.58 These two documents alone gave CAISI at least 21 distinct taskings.59
* * *
56 National Institute of Standards & Technology, "CAISI Evaluation of DeepSeek AI Models Finds Shortcomings and Risks," September 30, 2025.
57 The White House, "Memorandum of Understanding between the Government of the United States of America and the Government of the United Kingdom of Great Britain and Northern Ireland regarding the Technology Prosperity Deal," Presidential Memoranda, September 18, 2025.
58 U.S. Department of Commerce, "Statement from U.S. Secretary of Commerce Howard Lutnick on Transforming the U.S. AI Safety Institute into the Pro-Innovation, Pro-Science U.S. Center for AI Standards and Innovation," Press Release, June 3, 2025; The White House, "Winning the Race: America's AI Action Plan," July 2025.
59 Cole Salvador, Jack Crovitz, and Yusuf Mahmood, "Building AI Readiness in the U.S. Government," America First Policy Institute (AFPI), March 31, 2026.
* * *
Figure 2: U.S. Center for AI Standards and Innovation (CAISI) Funding Compared to Peer Institute
Caption: Committed funding (total, not annual) for the U.S. Center for AI Standards and Innovation (CAISI) lags far behind analogous institutes in other countries and remains 11 times smaller than the UK's equivalent.60
* * *
If CAISI is to help implement President Trump's AI agenda and foresee the threats posed by the CCP's AI ambitions, Congress must act to authorize it, fund it, and streamline its missions. In our report, we outline an America First vision for CAISI in which it serves the following roles: technical strike team, bridge between industry and government, frontier analysis unit, and technical standards organization.
Congress could consider allocating $50-100 million to CAISI annually. With a clarified mission and adequate funding, CAISI could become America's first line of defense against the CCP's illicit campaign to overtake America in AI.
* * *
60 Singapore Intercom Media Development Authority, "Digital Trust Centre designated as Singapore's AI Safety Institute," May 22, 2024; Government of Canada, "Canada launches Canadian Artificial Intelligence Safety Institute," November 12, 2024; UK Department for Science, Innovation, & Technology, "AI Opportunities Action Plan: One Year On," Policy Paper, January 29, 2026; Australian Assistant Minister for Science, Technology and the Digital Economy Andrew Charlton, "National AI Plan: Empowering all Australians," Press Release, December 2, 2025; U.S. Senate Committee for Appropriations, "BILL SUMMARY: Commerce, Justice, Science, and Related Agencies Fiscal Year 2024 Appropriations Bill," March 3, 2024.
* * *
The Bureau of Emerging Threats
The Bureau of Emerging Threats (ET) is an office in the State Department with similar promise and challenges to CAISI. It is "charged with anticipating and responding" to "U.S. adversaries' weaponization of advanced technology, including artificial intelligence."61 This mandate makes ET a natural candidate for analyzing the increasingly state-backed CCP campaign to acquire American AI. The Bureau could, for example, analyze trends in Chinese AI-enhanced military software like Chinese satellite intelligence reportedly being used by Iran to target U.S. forces.62 It might also analyze claims that AI threatens to undermine nuclear deterrence.63 To achieve these goals, Congress could authorize ET and provide it with direct funding. It could be mandated to deliver national security insights on emerging technology to key policymakers through regular reports.
Recommendations for the Committee
Congress has a key role to play in addressing and anticipating the CCP's offenses in the AI race. In this section, we recommend legislative actions that would help the U.S. Government counter the CCP's illicit campaign to acquire American AI technology. As a 501(c)(3), the America First Policy Institute does not support specific legislation.
Create an anti-distillation task force. Distillation attacks have become Chinese AI companies' preferred method of illegally extracting capabilities from American developers. Stopping distillation attacks would stop China's easiest path to free-riding on U.S. AI capabilities, but private American developers say that no single company can adequately defend against them alone. To combat distillation, Congress could mandate the NSA's AI Security Center coordinate with CAISI, the Bureau of Emerging Threats, other Intelligence Community offices, and industry to develop an anti-distillation task force. This task force could create and disseminate best practices for preventing distillation attacks, share threat intelligence, and conduct other relevant activities as appropriate.
Establish whistleblower protections. Congress could pass legislation to prohibit employment discrimination against whistleblowers reporting AI security vulnerabilities. Some security researchers at frontier AI companies complain that they are discouraged from informing policymakers about major security vulnerabilities.64 In particular, widespread severance and nondisclosure agreements (NDAs) at frontier labs prevent researchers from speaking out. This means that American policymakers and the national security community may not understand the seriousness of Chinese exfiltration and subversion operations until it is too late. As Senator Grassley has explained: "Today, too many people working in AI feel they're unable to speak up when they see something wrong. Whistleblowers are one of the best ways to ensure Congress keeps pace as the AI industry rapidly develops."65 Congress could consider passing legislation to protect security whistleblowers in the AI industry from retaliation by AI companies.
* * *
61 Shannon K. Kingston, "State Department launches effort to counter cyberattacks, AI risks from Iran, others," ABC News, March 23, 2026.
62 Henry Zartz, "Chinese AI satellite intelligence helping Iran target U.S. forces with 'incredible precision', analysts say," ABC News, April 6, 2026.
63 Jason Pruet, Anna Makanju, Jonathan Reiber, and Josh Achiam, "AI and International Security: Pathways of Impact and Key Uncertainties," OpenAI, February 6, 2026.
64 "OpenAI Employees Call for Protections to Speak Out on AI Risks," Bloomberg, June 4, 2024; Pranshu Verma, Kat Zakrzewski, & Nitasha Tiku, "OpenAI illegally barred staff from airing safety risks, whistleblowers say," Stars & Stripes, July 13, 2024; "OpenAI, Google DeepMind employees sign open
letter calling for whistle-blower protections to speak out on AI risks," South China Morning Post, June 5, 2024.
* * *
Establish basic security standards for frontier AI labs. Congress could pass legislation to empower the Department of War to institute minimum security standards for large AI developers in order to prevent Beijing from stealing or subverting American AI. The 2026 National Defense Authorization Act (NDAA) made a promising first step by directing the Pentagon to "develop a framework for the implementation of cybersecurity and physical security standards and best practices relating to covered artificial intelligence and machine learning technologies."66 However, the Pentagon has not yet published this framework, and the statute imposes no deadline by which the Department must do so. Given the seriousness of the Chinese security threat, Congress could build on this initiative by requiring the Department of War to develop and publish frontier AI lab security standards within a matter of months. Another weakness of the security standards authorized by the 2026 NDAA is that they apply only to AI labs that contract with the Pentagon. Given that some American frontier AI labs no longer sell AI tools to the Pentagon67, Congress could consider empowering the Department of War to impose security standards on all American AI labs regardless of their status as military contractors. These standards could include strong incident reporting requirements for attempted security breaches by nation-state adversaries.
Establish AI lab security red-teaming exercises. Congress could consider directing the NSA to lead redteaming exercises to uncover security vulnerabilities in the American AI development supply chain. NSA has deep expertise in offensive and defensive cyber operations, including classified threat vectors, so it is uniquely positioned to simulate the tactics of nation-state adversaries. Red-teaming exercises could target the full AI development process, from data scraping and filtering systems to training infrastructure to model weight storage facilities. They could also cover the full set of IP exfiltration and model poisoning attack vectors, including threats from compromised insiders. NSA's AI Security Center could collaborate with DHS and other parts of the intelligence community to ensure these exercises comprehensively simulate real-world attacks from nation-state adversaries. The exercises could be conducted in close partnership with frontier AI labs and on a fully voluntary basis.
* * *
65 U.S. Senate Committee on the Judiciary, "Grassley Introduces AI Whistleblower Protection Act," Press Release, May 15, 2025.
66 National Defense Authorization Act for Fiscal Year 2026, S. 1071, 119th Cong. (2025).
67 Cade Metz, Julian E. Barnes, and Sheera Frenkel, "Pentagon Officially Notifies Anthropic It Is a 'Supply Chain Risk'," New York Times, March 5, 2026.
* * *
Request a National Intelligence Estimate on Chinese AI. Several features of China's illicit campaign remain opaque: the extent of state support in some activities is unknown, the capability gains from different types of theft are difficult to quantify, and the intentions of Chinese leadership are unclear. In the 2026 National Defense Authorization Act, Congress required the Director of National Intelligence (DNI) to "produce a National Intelligence Estimate with respect to advancements by the [PRC] in biotechnology."68 Congress could ask DNI to produce a National Intelligence Estimate on advancements by the PRC in frontier AI. It would evaluate the country's drivers of AI progress, computing access and diversion, illicit campaign to acquire American AI, and other factors as appropriate.
Authorize and fund the Center for AI Standards and Innovation (CAISI). CAISI is a small, talent-dense technical office within the Department of Commerce that the Trump Administration and congressional leaders have recognized as a key asset in AI competition. But it lacks adequate staffing, funding, and authorization. Congress could allocate $50-100 million to CAISI annually and mandate it to act as a technical strike team, a bridge between industry and government, a frontier analysis unit, and a technical standards organization.
Authorize the Bureau of Emerging Threats (ET). ET is a new office in the Department of State with expertise in AI, military uplift, and international relations. It could complement CAISI by analyzing the strategic AI competition between the U.S. and the CCP from an international perspective and with an eye toward the future. For the Bureau to succeed, Congress could directly authorize it to deliver national security insights on emerging technology to key policymakers, such as through regular reports.
Conclusion
The United States and the Chinese Communist Party (CCP) are in a race to develop and deploy artificial intelligence. Overwhelming evidence shows that the CCP and its companies are using a suite of escalating actions to steal, subvert, and sabotage American AI. Unless we face this challenge with urgency and resolve, the CCP may succeed in its goal to achieve global AI leadership by 2030 -- an unacceptable outcome for American prosperity and power.
* * *
Biography
Yusuf Mahmood is Director of AI and Emerging Technology Policy at the America First Policy Institute (AFPI). He holds a J.D. from Harvard Law School and dual bachelor's degrees in economics and philosophy from the University of Maryland, College Park.
* * *
68 National Defense Authorization Act for Fiscal Year 2026, S. 1071, 119th Cong. (2025).
* * *
Original text and figures here: https://docs.house.gov/meetings/ZS/ZS00/20260416/119165/HHRG-119-ZS00-Wstate-MahmoodY-20260416.pdf
Amata-Cosponsored Bill to Suppress Illegal Fishing Advances
WASHINGTON, April 23 (Rep.) -- Del. Aumua Amata Radewagen, R-American Samoa, issued the following news release:* * *
Amata-Cosponsored Bill to Suppress Illegal Fishing Advances
*
Washington, D.C. - Congresswoman Uifa'atali Amata is welcoming advancement of a bill she cosponsors, the Fighting Foreign Illegal Seafood Harvests (FISH) Act, H.R. 3756, which cracks down on illegal, unreported and unregulated (IUU) fishing by strengthening inspections and penalties for vessels participating in the banned activity, especially the creation of an IUU vessel list.
"This bill addresses some key aspects ... Show Full Article WASHINGTON, April 23 (Rep.) -- Del. Aumua Amata Radewagen, R-American Samoa, issued the following news release: * * * Amata-Cosponsored Bill to Suppress Illegal Fishing Advances * Washington, D.C. - Congresswoman Uifa'atali Amata is welcoming advancement of a bill she cosponsors, the Fighting Foreign Illegal Seafood Harvests (FISH) Act, H.R. 3756, which cracks down on illegal, unreported and unregulated (IUU) fishing by strengthening inspections and penalties for vessels participating in the banned activity, especially the creation of an IUU vessel list. "This bill addresses some key aspectsof the fight against illegal fishing, and U.S. imports of illegally caught competitors' fish," said Congresswoman Aumua Amata. "Legislation like this can work in tandem with bolstered U.S. Coast Guard presence to reduce IUU fishing and protect U.S. food sourcing and commerce."
In 2019, IUU fishing was estimated to be involved in 11 percent of all U.S. fish imports, or 13 percent of imports caught at sea, known as marine capture, for a total of $2.4 billion in seafood in just one year, in this report by the U.S. International Trade Commission. The NOAA website notes, "The inherent nature of illegal, unreported, and unregulated fishing makes it difficult to accurately quantify the full global economic impacts resulting from these activities."
The bipartisan bill was one of eight advanced by the House Committee on Natural Resources (HNR). The FISH Act clarifies key definitions for enforcement; states U.S. policy; establishes an IUU vessel list and procedures for placing an offending vessel on that list; specifies other enforcement provisions, sanctions, penalties and exceptions; amends the Interagency Working Group on IUU Fishing; optimizes data collecting, sharing and analysis; takes steps to prevent importation from countries with forced labor; encourages investment and technical assistance for the U.S. fisheries sector, and provides for reports to Congress.
HNR Chairman Bruce Westerman, R-Ark., said, "House Committee on Natural Resources members are dedicated to enacting commonsense policies to support rural communities across America. Today's markup advanced legislation that will allow for more resource development, implement land exchanges for critical water infrastructure and the consolidation of culturally significant land, combat illegal fishing activities and strengthen historic battlefield conservation. I look forward to working with the sponsors to advance these bills through the House."
H.R. 3756 is sponsored by Rep. Dan Crenshaw, R-TX; with Rep. Seth Magaziner, D-RI; Rep. Nicholas Begich, R-Alaska; Rep. Aumua Amata, American Samoa; Rep. Nancy Mace, R-SC; and Rep. Eugene Vindman, D-VA. The House bill has companion legislation in the U.S. Senate, introduced by Sen. Dan Sullivan, R-Alaska.
Issues : Fisheries
***
Original text here: https://radewagen.house.gov/media-center/press-releases/amata-cosponsored-bill-suppress-illegal-fishing-advances
ALSOBROOKS: RFK JR. STILL WILL NOT ADMIT HE SAID "EVERY BLACK KID SHOULD BE RE-PARENTED"
WASHINGTON, April 23 -- Sen. Angela Alsobrooks, D-Maryland, issued the following news release on April 22, 2026:* * *
ALSOBROOKS: RFK JR. STILL WILL NOT ADMIT HE SAID "EVERY BLACK KID SHOULD BE RE-PARENTED"
Senator Alsobrooks (D-Md.) questioned RFK Jr. during the Health, Education, Labor and Pensions Committee's "Hearing on Fiscal Year 2027 Department of Health and Human Services Budget."
Last week during a House Ways and Means Committee hearing, Rep. Terri Sewell (Ala.-07) questioned Health and Human Services Secretary RFK Jr. about his previous comments stating, "Every Black kid is now ... Show Full Article WASHINGTON, April 23 -- Sen. Angela Alsobrooks, D-Maryland, issued the following news release on April 22, 2026: * * * ALSOBROOKS: RFK JR. STILL WILL NOT ADMIT HE SAID "EVERY BLACK KID SHOULD BE RE-PARENTED" Senator Alsobrooks (D-Md.) questioned RFK Jr. during the Health, Education, Labor and Pensions Committee's "Hearing on Fiscal Year 2027 Department of Health and Human Services Budget." Last week during a House Ways and Means Committee hearing, Rep. Terri Sewell (Ala.-07) questioned Health and Human Services Secretary RFK Jr. about his previous comments stating, "Every Black kid is nowjust standard put on Adderall, SSRIs, benzos, which are known to induce violence. And those kids are going to have a chance to go somewhere and get re-parented -- to live in a community where there'll be no cellphones, no screens."
RFK Jr. denied saying it, and today, Senator Alsobrooks confronted him with a transcript and recording.
Senator Alsobrooks also questioned RFK Jr. on the role of his friend and close associate, Aaron Siri, an anti-vaccine advocate and lawyer who has been working to push RFK Jr. to pave the way for HHS to falsely connect vaccines with autism, and reduce access to life-saving vaccines for American children. When questioned, RFK Jr. couldn't answer who is actually deciding policies at HHS - the White House or Aaron Siri.
See below for one of their exchanges:
Senator Alsobrooks: My next question. Now, you've had some trouble with the truth. I've seen it myself during your appearances before Congress. We all saw during your exchange last week with Congresswoman Sewell. Can you admit today that you said every Black kid can get re-parented on a wellness farm? Can you admit that you said that?
RFK Jr.: Get re-parented on a wellness farm?
Senator Alsobrooks: Well, let me read exactly what you said, you said, "Every Black kid is now just standard put on Adderall, SSRIs, benzos, which are known to induce violence. And those kids are going to have a chance to go somewhere and get re-parented -- to live in a community where there'll be no cellphones, no screens."
The whole transcript was even worse. You said that if you could, you would send every Black kid, your words not mine, to go live on farms and work.
RFK Jr.: I would have to hear that recording. Because I have no memory of saying anything like that.
Senator Alsobrooks: Well, I actually have the recording that I can give to you. But it is absolutely what you said. And if you want me to play it, I can play it.
RFK Jr.: If you ask me what my opinion is, I do not believe that every Black kid should be re-parented on a wellness farm or whatever, and I have never believed that.
Senator Alsobrooks: Well, you said it, sir. I have the video here.
RFK Jr.: I'm telling you I don't believe it. That's not my vision for our country.
Senator Alsobrooks: Well, I'm glad because it was ignorant to say. It was dangerous and it was irresponsible.
RFK Jr.: Well, if I said it, I apologize, but I'd have to see the transcript.
* * *
Original text here: https://www.alsobrooks.senate.gov/news/press-releases/alsobrooks-rfk-jr-still-will-not-admit-he-said-every-black-kid-should-be-re-parented/
Alsobrooks 'No' on Kevin Warsh
WASHINGTON, April 23 -- Sen. Angela Alsobrooks, D-Maryland, issued the following statement on April 21, 2026:* * *
Alsobrooks 'NO' on Kevin Warsh
Senator Angela Alsobrooks (D-Md.) released the following statement after questioning President Trump's nominee for Chairman of the Federal Reserve, Kevin Warsh.
"I am a 'no' on Kevin Warsh for Chairman of the Federal Reserve. He could not answer my very simple questions regarding the independence of the Federal Reserve amid the politically motivated investigations into Federal Reserve Chair Jerome Powell and Governor Lisa Cook. Mr. Warsh did not ... Show Full Article WASHINGTON, April 23 -- Sen. Angela Alsobrooks, D-Maryland, issued the following statement on April 21, 2026: * * * Alsobrooks 'NO' on Kevin Warsh Senator Angela Alsobrooks (D-Md.) released the following statement after questioning President Trump's nominee for Chairman of the Federal Reserve, Kevin Warsh. "I am a 'no' on Kevin Warsh for Chairman of the Federal Reserve. He could not answer my very simple questions regarding the independence of the Federal Reserve amid the politically motivated investigations into Federal Reserve Chair Jerome Powell and Governor Lisa Cook. Mr. Warsh did notprovide me or my constituents with the certainty that he would be an independent voice overseeing our economy--so I cannot vote to confirm him for this position," said Senator Alsobrooks.
See below for one of the exchanges:
Senator Alsobrooks: In August of 2025 President Trump tried to fire Fed Governor Lisa Cook for unsubstantiated allegations widely believed to be politically motivated. Thankfully, both a federal district court and the D.C. Circuit Court of Appeals ruled in favor of Governor Cook, preventing her from being fired while the investigation proceeds. This is a landmark test for independence. Now, Supreme Court Justice Brett Kavanaugh said allowing Cook's firing to go forward would weaken if not shatter the independence of the Federal Reserve. So I would like to ask you: Will you commit to defending Governor Cook's tenure as Chairman Powell has done?
Kevin Warsh: Senator, it was a pleasure to meet you in your office and spend time with you. As I said to you then, which I will repeat now to the broader committee, if I stand for anything, it's the Fed should stay in its lane. As I understand that matter, it's pending before the United States Supreme Court. I think it is inappropriate for me to weigh in on that, especially because in the event that I am confirmed, I could be a party to that matter.
Senator Alsobrooks: Well, I disagree with you, Mr. Warsh. I disagree with you. This is your future colleague who is confirmed by both this committee and this Senate to serve her country. Chairman Powell has defended her tenure.
* * *
Original text here: https://www.alsobrooks.senate.gov/news/press-releases/alsobrooks-no-on-kevin-warsh/
Alford Commends Appropriations Committee Passage of FY27 Financial Services & General Government Funding Bill
WASHINGTON, April 23 -- Rep. Mark Alford, R-Missouri, issued the following news release:* * *
Alford Commends Appropriations Committee Passage of FY27 Financial Services & General Government Funding Bill
*
Today, Congressman Mark Alford (MO-04) issued the following statement after the House Appropriations Committee advanced the Fiscal Year 2027 Financial Services and General Government (FY27 FSGG) funding bill.
"For the second time this week, the House Appropriations Committee is advancing Fiscal Year 2027 funding ahead of schedule. This fiscally responsible legislation cuts spending by ... Show Full Article WASHINGTON, April 23 -- Rep. Mark Alford, R-Missouri, issued the following news release: * * * Alford Commends Appropriations Committee Passage of FY27 Financial Services & General Government Funding Bill * Today, Congressman Mark Alford (MO-04) issued the following statement after the House Appropriations Committee advanced the Fiscal Year 2027 Financial Services and General Government (FY27 FSGG) funding bill. "For the second time this week, the House Appropriations Committee is advancing Fiscal Year 2027 funding ahead of schedule. This fiscally responsible legislation cuts spending bynearly $1 billion from last year-while advancing President Trump's America First agenda by eliminating wasteful spending and ending divisive Biden-era policies," said Congressman Alford. "The bill strengthens cybersecurity and IT modernization across government, protects consumer freedoms, prohibits the creation of a Central Bank Digital Currency, safeguards election integrity, and defends First Amendment rights.
"I'm incredibly proud this bill includes our Strategic Assets Protection Act to protect Whiteman AFB and other nuclear triad facilities from foreign adversaries. I'm pleased to help advance this commonsense legislation that delivers a more efficient, accountable, and secure government for the hardworking families and taxpayers of Missouri and across our nation."
Fourth District Priorities Included in the FY27 FSGG Bill:
* Rep. Alford's Strategic Assets Protection Act, which will ensure sensitive national security sites associated with our nuclear triad, including Whiteman Air Force Base, are protected from foreign adversaries by directing the Committee on Foreign Investment in the United States (CFIUS) to review all foreign transactions involving real estate, businesses, critical infrastructure, or assets surrounding these facilities since January 1, 2017.
* Watch Rep. Alford's remarks in committee on this provision here.
* A provision requiring the Inspector General for the USPS to investigate and issue a report on the abysmal service quality and deliver performance plaguing the Kansas City metro area.
* Language encouraging the Securities and Exchange Commission to allow insurance companies to exclude woke, left-wing Environmental, Social, and Governance (ESG) proposals from their proxy materials.
* Several provisions encouraging the agencies funded under the bill, including the IRS, Office of Personnel Management, and Treasury Department, to utilize innovative technology to ensure the sensitive personal information of taxpayers and federal employees are adequately protected.
Background on the FY27 FSGG Bill:
Strengthening National Security Infrastructure
* New funding for CFIUS, TFI, FinCEN and Cybersecurity
* Level funding for High Intensity Drug Trafficking Areas (HIDTA) program and other drug control programs within the Office of National Drug Control Policy (ONCDP)
Leveraging New Technology
* This includes prioritizing AI, machine learning, and other modernization tools to both strengthen IT platforms and ensure the federal government runs effectively and efficiently
Prioritizing Fiscal Responsibility
* This includes shrinking the size of agencies to reflect current workforce trends
* Aligned agencies funding with the personnel cuts the President has made
* Facilitating the disposal of dilapidated buildings while ensuring maintenance resources are available
* Cuts IRS enforcement by 28% or $1.4 billion
Cracking down on Fraud
* Fully funds IGs and PRAC to stop waste, fraud, abuse as well as improper and fraudulent payments
* Codifying Executive Orders supporting efforts to detect and stop improper payments
Federal Judiciary Security
* Fully funds security request for Supreme Court Justices and all federal Judges
Pro Life Protections:
* Retains prohibitions on sex denying care and DC Abortion Riders
***
Original text here: https://alford.house.gov/news/documentsingle.aspx?DocumentID=1559
After Pressure From Senator Gillibrand, Trump Administration To Release Energy Assistance Funding
WASHINGTON, April 23 -- Sen. Kirsten Gillibrand, D-New York, issued the following news release:* * *
After Pressure From Senator Gillibrand, Trump Administration To Release Energy Assistance Funding
*
The announcement comes as relief for working families across New York and the country who are facing sticker shock from their eye-popping gas and utility bills.
NY will receive an additional $40.3 million in federal funding to help them afford home energy bills.
U.S. Senator Kirsten Gillibrand (D-NY), a member of the Senate Appropriations Committee, has successfully pushed the Trump administration ... Show Full Article WASHINGTON, April 23 -- Sen. Kirsten Gillibrand, D-New York, issued the following news release: * * * After Pressure From Senator Gillibrand, Trump Administration To Release Energy Assistance Funding * The announcement comes as relief for working families across New York and the country who are facing sticker shock from their eye-popping gas and utility bills. NY will receive an additional $40.3 million in federal funding to help them afford home energy bills. U.S. Senator Kirsten Gillibrand (D-NY), a member of the Senate Appropriations Committee, has successfully pushed the Trump administrationto release over $400 million to help New York families afford their energy bills.
The announcement comes as relief for working families across New York and the country who are facing sticker shock from their eye-popping gas and utility bills. According to the National Energy Assistance Directors Association, roughly 1 out of 6 U.S. households ( 21.5 million nationwide ) are behind on their energy bills in 2026.
Earlier this month, Senator Gillibrand and a bipartisan group of senators demanded that the Trump administration release over $400 million in remaining Low Income Home Energy Assistance Program (LIHEAP) funding that Congress delivered in fiscal year (FY) 2026.
Following this pressure from Senator Gillibrand, New York will receive an additional $40,356,315 in LIHEAP funds to go along with the more than $360 million in federal funding that the state previously received. As a result, New York will net a total of $401.2 million in LIHEAP aid this year to help New Yorkers afford their home utility bills. This means a $20 million increase over the previous year.
"LIHEAP is a commonsense, bipartisan program," said Senator Gillibrand. "In the coldest and hottest months of the year, it lowers the cost of living and saves lives. President Trump is already causing fuel and gas prices to skyrocket due to his reckless choice to start a war in Iran. He continued to make life harder for working Americans by stalling hundreds of millions of dollars in federal funding that Congress already allocated to LIHEAP from reaching families in need. Finally releasing these funds is a welcome step, and I will continue pushing back on the administration's recent budget proposal that eliminates LIHEAP in order to fund an unauthorized war of choice."
LIHEAP is a federally funded program that helps low-income households pay utility bills, address energy crises, and lower costs by improving home energy efficiency through weatherization. Over the last year, nearly six million households nationwide - including about 1.5 million in New York -received LIHEAP assistance to help them lead a healthier, more stable life and avoid having their utilities shutoff or having to make tradeoffs such as skipping medical care or meals or turning to unfair payday loans that lead to a cycle of unending debt.
Earlier this month, the Trump administration released a $2.2 trillion Fiscal Year 2027 budget proposal that would completely eliminate all LIHEAP funding and slash ten percent of other domestic programs across the board while increasing defense spending. And last year, the Trump administration fired the entire LIHEAP program staff at the U.S. Department of Health and Human Services (HHS).
In addition to Gillibrand, the April 15 letter was signed by U.S. Senators Jack Reed (D-RI), Susan Collins (R-ME), Chuck Schumer (D-NY), Richard Blumenthal (D-CT), Charles E. Grassley (R-IA), Angus S. King, Jr. (I-ME), Dan Sullivan (R-AK), Lisa Murkowski (R-AK), Susan Collins (R-ME), Elissa Slotkin (D-MI), Sheldon Whitehouse (D-RI), Jack Reed (D-RI), Tammy Baldwin (D-WI), Ben Ray Lujan (D-NM), Lisa Blunt Rochester (D-DE), Jeff Merkley (D-OR), Mark Warner (D-VA), Ruben Gallego (D-AZ), Jacky Rosen (D-NV), Mazie K. Hirono (D-HI), Michael Bennet (D-CO), Chris Coons (D-DE), Elizabeth Warren (D-MA), Raphael Warnock (D-GA), Chris Van Hollen (D-MD), Angela D. Alsobrooks (D-MD), Edward J. Markey (D-MA), Tim Kaine (D-VA), Catherine Cortez Masto (D-NM), Ron Wyden (D-OR) Adam Schiff (D-CA), Peter Welch, Richard Durbin (D-IL), John Fetterman (D-PA), Tina Smith (D-MN), Jeanne Shaheen (D-NH), Amy Klobuchar (D-MN), Tammy Duckworth (D-IL), Mark Kelly (D-AZ) Bernie Sanders (I-VT), and Cory Booker (D-NJ).
The text of the letter can be found here.
***
Original text here: https://www.gillibrand.senate.gov/news/press/release/after-pressure-from-senator-gillibrand-trump-administration-to-release-energy-assistance-funding/