2026 U.S. Government Activity on Artificial Intelligence

By Khloe McDonough

Recent federal actions and proceedings in March 2026 highlight ongoing efforts by the U.S. government to shape artificial intelligence (AI) policy, address national security concerns related to foreign AI technologies, and manage relationships with domestic AI developers.

White House Release of the National AI Legislative Framework

 On March 20, 2026, the White House released a National AI Legislative Framework[1] outlining policy recommendations for Congress.  The framework is intended to support the development of a unified federal approach to AI regulation and address six key objectives: (1) protecting children and empowering parents, (2) safeguarding and strengthening communities, (3) respecting intellectual property rights (IP) and creators, (4) preventing censorship and protecting free speech, (5) enabling innovation, (6) and educating Americans for an AI-ready workforce[2].

The framework indicates that the use of copyrighted material in AI model training constitutes infringement and should continue to be addressed through existing legal processes. However, the framework does not propose new statutory standards. The framework further emphasizes federal preemption of certain state AI laws, particularly in states where AI regulations are inconsistent[3].

Congressional Hearing on DeepSeek and Unitree Robotics

On March 17, 2026, the House Committee on Homeland Security held a hearing to examine the national security risks associated with AI and robotics technologies developed by companies based in the People’s Republic of China, including DeepSeek and Unitree Robotics. The hearing focused on potential risks related to data access, technology transfer, and the use of foreign-developed AI systems in sensitive contexts. The hearing also addressed broader strategic issues, including U.S. technological leadership and reducing reliance on foreign AI technologies.

A central issue discussed was the risk of exposure to sensitive data and critical systems using foreign-developed technologies. Witnesses noted that such systems could introduce cybersecurity vulnerabilities, enable data collection, or create dependencies on foreign suppliers[4]. Furthermore, the hearing addressed issues related to potential use for reverse engineering or reliance on outputs from existing systems[5]. Lawmakers highlighted the dual-use nature of AI and robotics technologies, noting that these technologies can serve both commercial and military defense-related purposes.

To address these issues, policymakers and witnesses discussed several potential responses, including strengthening export controls, limiting the use of foreign technologies, increasing domestic investment in AI and robotics, and enhancing supply chain security measures.

Developments Involving the U.S. Government and Anthropic

Recent developments involving Anthropic and the Trump administration highlight the ongoing dispute over the use of privately owned AI systems in government. When Anthropic declined to agree to certain U.S. government contract terms that would have permitted the Department of Defense access to their Claude AI model, including applications to unrestricted autonomous operations and surveillance, the Pentagon designated Anthropic as a “supply chain risk” to limit and prohibit the use of the company’s technology in the government’s systems[6]. In response, the Trump administration had begun a six-month phase-out of Anthropic’s software since February[7].

Anthropic subsequently filed legal challenges in federal court[8], arguing that the designation was not based on legitimate security concerns, but was instead retaliatory in nature. Anthropic raised constitutional claims, including alleged violations of due process and free speech protections, asserting that the restrictions will result in significant financial and operational harm to the company[9]. In response, a federal judge of Northern California[10] issued a temporary order blocking enforcement of the designation for the duration of the case, with the court indicating there were sufficient questions about the basis of the government’s actions to justify interim relief[11].

As of late April 2026, the Trump administration has begun to ease its restrictions on federal use of Anthropic software. The shift in the administration follows a recent White House meeting between administration officials and Anthropic’s CEO, where discussions focused largely on the company’s advanced AI program, Mythos, rather than the ongoing legal and regulatory challenges facing Anthropic[12]. Mythos has drawn attention for its sophisticated ability to identify software vulnerabilities, capabilities that carry significant implications for national cybersecurity. While such tools could strengthen defensive infrastructure, they also raise concerns about potential misuse.

Despite the optimistic conversations that occurred in the White House meeting, tensions remain. The United States Department of Defense has not withdrawn its charges, and the administration is still at a pause in its court case against Anthropic[13]. At the same time, the U.S. Department of Commerce Center for AI Standards and Innovation has initiated testing of Mythos, and agencies such as the United States Department of the Treasury have expressed interest in obtaining access[14], reflecting how the federal government is balancing contesting Anthropic’s practices while considering the strategic advantages of its technology.

References

[1] President Donald J. Trump Unveils National AI Legislative Framework, The White House (Mar. 20, 2026), https://www.whitehouse.gov/releases/2026/03/president-donald-j-trump-unveils-national-ai-legislative-framework/

[2] Id.

[3] Id.

[4] DeepSeek and Unitree Robotics: Examining the National Security Risks of PRC Artificial intelligence, Robotics, and Autonomous Technologies and Building a Secure U.S. Technology Base, Hearing before the Subcomm. on Cybersecurity & infrastructure Prot. of the H. Comm. on Homeland Sec., 119th Cong. (Mar. 17, 2026), https://homeland.house.gov/hearing/deepseek-and-unitree-robotics-examining-the-national-security-risks-of-prc-artificial-intelligence-robotics-and-autonomous-technologies-and-building-a-secure-u-s-technology-base/

[5] Id.

[6] Karen Freifeld & Alexander Alper, Anthropic Dispute with Pentagon Over AI Safeguards Escalates, Reuters (Mar. 11, 2026).  

[7] Id.

[8] Anthropic PBC v. U.S. Department of War, No. 3:26-cv-01996-RFL, Compl. (N.D. Cal. Mar. 9, 2026).

[9] Anthropic Files Federal Lawsuit Challenging U.S. Restrictions, PBS NewsHour (Mar. 2026).

[10] Anthropic PBC v. U.S. Department of War, No. 3:26-cv-01996-RFL, Order Granting Preliminary Injunction, slip op. (N.D. Cal. Mar. 9, 2026).

[11] David Shepardson, U.S. Judge Blocks Pentagon Blacklisting of Anthropic, Reuters (Mar. 26, 2026).

[12] Brendan Bordelon, Trump Picked a Fight with Anthropic. Now the Administration is Backing Off, POLITICO (Apr. 23, 2026), https://www.politico.com/news/2026/04/23/trump-picked-a-fight-with-anthropic-now-the-administration-is-backing-off-00889241?utm_campaign=mil-ebb&utm_medium=email&utm_source=sailthru

[13] Id.

[14] Id.