The White House Looks to the Future of Artificial Intelligence in the US
Though AI legislation in the US remains nascent, the Obama, Trump, and Biden administrations have all voiced opinions on what the future of AI should look like in the US. The Trump administration’s AI-directed communications recently culminated in the release of “A National Policy Framework for Artificial Intelligence” [1] (the Framework). The Framework is a continuation of the December 2025 Executive Order “Ensuring a National Policy Framework for Artificial Intelligence” [2] (the Executive Order) and advocates for many of the same priorities. Unlike the Executive Order, however, the Framework includes clear Congressional calls to action, addresses intellectual property rights, and proposes strategies for enabling innovation.
Overview of the Framework
The Framework outlines seven key priorities:
1. Protecting Children and Empowering Parents
The Framework instructs Congress to provide parents and guardians with tools that enable management of their children’s digital environment and to establish and affirm requirements of AI platforms relating to privacy protection of minors. Congress is advised against preempting states from enforcing generally applicable laws protecting children.
2. Safeguarding and Strengthening American Communities
The Framework encourages acceleration of AI infrastructure development and AI resource usage while advising Congress to prevent harmful impacts such as residential electricity cost increases and AI impersonation scams.
3. Respecting Intellectual Property Rights and Supporting Creators
The Framework states that “the Administration believes that training of AI models on copyrighted material does not violate copyright laws” but supports the Courts in resolving the issue. The Framework encourages Congress to build on any court decisions, e.g., by enabling licensing frameworks for rights holders and protecting against unauthorized use of AI-generated reproductions, while protecting First Amendment rights.
4. Preventing Censorship and Protecting Free Speech
The Framework instructs Congress to prevent the US government from coercing technology providers into regulating content based on partisan or ideological agendas. The Framework further instructs Congress to provide a means for Americans to seek redress from the US government in cases of government censorship or overreach on AI platforms.
5. Enabling Innovation and Ensuring American AI Dominance
The Framework insists on federal cooperation with AI initiatives, advising Congress to establish regulatory sandboxes for AI applications, make federal datasets accessible in AI-ready formats, and support existing regulatory bodies in developing and deploying sector-specific AI applications. The Framework explicitly advises Congress against creating “any new federal rulemaking body to regulate AI.”
6. Educating Americans and Developing an AI-Ready Workforce
The Framework advises Congress to use non-regulatory methods to incorporate AI training into existing youth development and skills training programs and to study AI-driven trends in the workforce.
7. Establishing a Federal Policy Framework, Preempting Cumbersome State AI Laws
The Framework encourages the creation of a national AI standard that preempts state AI laws, specifically those relating to AI development, Americans’ use of AI for lawful activities, and unlawful conduct by third parties using AI models. However, the Framework also notes the importance of respecting federalism, specifically in the areas of traditional police powers retained by the states, state zoning laws relating to placement of AI infrastructure, and requirements governing the use of AI by each state.
The Framework’s emphasis on preempting state legislation was expected based on other White House communications and is described in the Framework as motivated by a desire to prevent a “fragmented patchwork of state regulations that would hinder our national competitiveness.” But how fragmented are state regulations and how exactly could this fragmentation affect national competitiveness?
The Framework versus Regulatory Efforts at the State Level
In an August 2025 Brookings article on “How different states are approaching AI” [3], researchers found that a total of 260 state AI bills were introduced in the 2025 legislative session and twenty-two bills were passed. Though all fifty states, Puerto Rico, the Virgin Islands, and Washington, D.C. introduced AI-related legislation, very few states attempted to pass comprehensive AI governance bills and many of these bills were abandoned or significantly narrowed in scope [4]. In fact, Colorado’s “Consumer Protections for Artificial Intelligence” bill is the only comprehensive bill to have passed and has yet to be implemented.
Accordingly, the vast majority of state-level legislative activity has been category-specific. For example, there has been bipartisan support for initiatives relating to regulating non-consensual intimate imagery (NCII) and child sexual abuse material (CSAM), the use of AI in elections, generative AI transparency, automated decision-making technology (ADMT) and high-risk AI, state government use of AI, the impact of AI on employment, and the use of AI systems in healthcare decisions. The initiatives relating to regulating NCII, CSAM, and the state government use of AI are the only categories explicitly protected against preemption according to the Framework.
Despite an overlap in categories of legislation proposed in both left- and right-leaning states, trends have emerged in the conditions most amenable to AI policy. In a 2026 Brookings article on “Why AI policy thrives in some states and fades in other” [5], it was observed that states with the most AI bills introduced tend to be Democrat-leaning states with younger populations (e.g., Nevada and Virginia) and/or high-per capita income states led by Democratic governors (e.g., Massachusetts and Connecticut). Similarly, states with an absence of AI bills introduced tend to have a Republican-leaning electorate (e.g., Arkansas, Indiana, Missouri, and Wyoming), older populations and lower per capita income (e.g., Hawaii, Maine, Pennsylvania, and Wisconsin), and/or states with older populations led by Republican governors (e.g., New Hampshire and West Virginia). These trends are generally reflected in areas of individual protection and transparency, with a Republican party base and worse economic standing corresponding with lower protection of individuals and fewer (or no) transparency requirements.
Due to the wide variety of approaches that states are taking when it comes to AI regulations and the significant inertia associated with getting any new AI regulation passed, it may be fair to characterize the current state AI regulations as a “fragmented patchwork.”
The Framework versus Regulatory Efforts Outside the US
Turning to the question of national competitiveness, the Framework indicates that state-level regulations may increase barriers to innovation and reduce access to data, thereby preventing American AI dominance. The Framework’s recommended state-level legislative preemptions work in tandem with its cautions against broad federal-level AI regulations, and security concerns are presented as the motivator for any potential oversight of AI systems.
The US approach to AI regulations may be contrasted with the regulatory approaches taken by the European Union (EU) and China, often viewed as the US’s major competitors in the race to AI dominance. The EU, primarily through the EU AI ACT adopted in 2024, has enacted a broad regulatory approach to AI systems by classifying these systems into risk categories with associated requirements. Though many legal experts agree with the EU’s human-centric approach to AI regulation, there are concerns that the extent of regulation may prevent companies from developing and deploying AI (i.e., slowing innovation) [6]. China has taken a different approach, enacting targeted AI laws with broader AI regulation in development. China’s approach appears to be an attempt at balancing corporate and individual interests, though its success remains to be seen [7].
Conclusion
The Framework reaffirms the current Administration’s decentralized approach to AI regulation while acknowledging that certain steps should be taken by Congress to protect individuals, especially minors, from potential harms associated with AI systems. Much of the Framework’s guidance is non-regulatory, focused on ways Congress can support individuals, families, and companies in navigating AI systems. Furthermore, parts of the Framework are anti-regulation, instructing Congress to refrain from creating a new federal rulemaking body to regulate AI and preempt certain state regulations, some of which are currently enacted.
Those working in the field of intellectual property may appreciate the nods to intellectual property rights and innovation in the Framework while also feeling some disappointment at the uncertainty that remains on navigating AI-related matters. Specifically, the Framework takes the controversial position that “training of AI models on copyrighted material does not violate copyright laws,” though it defers to the Courts on settling the issue. With over 100 copyright suits against AI companies pending in the US [8], there is a great deal of uncertainty that remains.
References
[1] The White House, A National Policy Framework for Artificial Intelligence (2026)
[2] Executive Order, Ensuring a National Policy Framework for Artificial Intelligence (2025)
[3] Josie Stewart, Nicol Turner Lee, et al., How different states are approaching AI, Brookings (2025)
[4] National Conference of State Legislatures, Artificial Intelligence 2025 Legislation, https://www.ncsl.org/technology-and-communication/artificial-intelligence-2025-legislation (last visited Apr 20, 2026)
[5] Gregory S. Dawson, James S. Denford, et al., Why AI policy thrives in some states and fades in others, Brookings (2026)
[6] Laurie Harris, Regulating Artificial Intelligence: U.S. and International Approaches and considerations for Congress (2025)
[7] Yue Zhu et al., China’s emerging regulation toward an open future for AI, Science 390,132-135(2025). DOI:10.1126/science.ady7922
[8] Latest world map of copyright suits v. AI companies (April 5, 2026), Chat GPT Is Eating the World (2026), https://chatgptiseatingtheworld.com/2026/04/05/latest-world-map-of-copyright-suits-v-ai-companies-april-5-2026/ (last visited Apr 20, 2026)
