Evolving AI Regulation: Federal & State Policy Frameworks

The dawn of 2026 marks a pivotal period in the landscape of artificial intelligence (AI) governance, as governments globally grapple with the intricate challenge of fostering innovation while simultaneously erecting robust safeguards. The United States, in particular, is witnessing a flurry of activity, with both federal and state entities racing to establish comprehensive AI Regulation and policy frameworks. This concerted effort highlights a growing recognition that unstructured AI development, while revolutionary, carries inherent risks that demand structured oversight. The tension between a unified national standard and diverse state-level approaches is defining the current regulatory discourse, shaping the future of AI’s integration into society.

The Federal Blueprint: White House National Policy Framework for AI

On March 20, 2026, the White House unveiled its National Policy Framework for Artificial Intelligence, a landmark document outlining legislative recommendations designed to guide Congress toward a coherent, nationally unified approach to AI governance. This framework, while non-binding, is poised to significantly influence federal AI legislation in the coming months and years.

Seven Pillars of Federal AI Governance

The White House’s comprehensive vision for federal AI policy is anchored in seven thematic pillars, balancing innovation, competitiveness, and national security with targeted safeguards for individuals and communities:

1. Protecting Children and Empowering Parents: A core focus is on safeguarding minors from AI-related risks. The framework recommends that Congress establish privacy protections and age-verification requirements for AI services likely to be accessed by children. This includes providing parents with tools to manage their children’s privacy settings, screen time, and content exposure. Furthermore, it urges AI platforms to implement features that reduce the risks of sexual exploitation and self-harm to minors and to continue enforcing prohibitions on nonconsensual disclosures of intimate depictions. Notably, any federal legislation should not preempt states from enforcing their generally applicable laws protecting children, such as prohibitions on child sexual abuse material.
2. AI Infrastructure and Small Business Support: This pillar aims to facilitate the growth of AI infrastructure while protecting communities from associated harms. Recommendations include streamlining federal permitting for the construction and operation of AI facilities. It also supports AI developers’ ability to develop on-site power generation, simultaneously protecting residential ratepayers from increased energy costs related to AI data centers. The framework also stresses augmenting federal law enforcement efforts against AI-enabled impersonation scams and fraud targeting vulnerable populations. Moreover, it supports small businesses with grants, tax incentives, and technical assistance for AI adoption.
3. Intellectual Property: The framework emphasizes protecting creators’ works and identities while preserving innovation. It recommends that Congress provide protections for individuals affected by the unauthorized distribution or commercial use of AI-generated digital replicas of their voice, likeness, or other identifiable attributes, while exempting parody, satire, news reporting, and other expressive works protected by the First Amendment. The framework also recommends that Congress consider enabling collective licensing frameworks that would allow rights holders to negotiate compensation from AI providers, though it defers to courts on unsettled questions of copyright law like fair use.
4. Censorship and Free Speech: Echoing concerns regarding compelled speech and government overreach, the framework stresses that AI should not be used by government actors to suppress lawful expression. It calls on Congress to prevent government coercion of platforms and AI providers and to provide redress where censorship-related harms stem from government action, while avoiding regulation of private content moderation decisions.
5. Enabling Innovation: To remove barriers to innovation and accelerate AI deployment, the framework recommends that Congress establish regulatory “sandboxes” for AI applications to support experimentation. It also advocates for making federal datasets more accessible for AI training in “AI-ready formats”. Significantly, the framework recommends against creating any new federal rulemaking body to regulate AI, calling instead for AI to be governed through existing regulatory agencies with subject-matter expertise and industry-led standards.
6. Workforce Preparation: This pillar addresses the impact of AI on the American workforce. The framework recommends integrating AI training into existing education and workforce development programs through non-regulatory methods. It also calls for expanding federal efforts to study trends in AI-driven workforce realignment to inform supportive policies and bolster capabilities at land-grant institutions to provide technical assistance, launch demonstration projects, and develop youth-centered AI programs.
7. Preemption of State AI Laws: Perhaps the most consequential aspect, the framework recommends that Congress broadly preempt state AI laws that “impose undue burdens.” The stated goal is to establish a single, minimally burdensome national standard rather than a fragmented patchwork of fifty discordant ones, which is seen as hindering innovation and US competitiveness. However, it carves out several categories of state law from preemption, including generally applicable laws protecting children, fraud, and consumer protection, state zoning authority, and states’ own uses of AI for law enforcement or other public services.

State-Level Momentum: A Patchwork of Proactive Legislation

Despite the federal push for preemption, several states are actively advancing their own AI-related legislation, creating a dynamic and, at times, conflicting regulatory landscape. This state-level activity underscores the urgent need for AI Regulation tailored to specific local concerns and priorities.

Tennessee’s Proactive Stance on Mental Health AI

On April 1, 2026, Tennessee Governor Bill Lee signed SB 1580 into law, taking effect on July 1, 2026. This legislation prohibits individuals or entities that develop or deploy AI systems from advertising or representing to the public that such systems are, or are able to act as, a qualified mental health professional. The law defines AI as “models and systems capable of performing functions generally associated with human intelligence, including reasoning and learning.” A violation of SB 1580 constitutes an unfair or deceptive act under the Tennessee Consumer Protection Act of 1977, carrying a civil penalty of up to $5,000 per violation and establishing a private right of action for affected parties. This targeted approach aims to ensure that mental health services remain firmly within the domain of human professionals, responding to concerns about AI companion use and potential negative impacts on mental health.

Georgia’s Chatbot Disclosure and Child Safety Measures

Georgia has also made significant strides with the approval of SB 540, a chatbot disclosure and child safety bill, which was sent to Governor Brian Kemp’s desk on April 6, 2026. This bill, drawing national attention for its breadth and lack of industry exemptions, mandates several crucial requirements for operators of conversational AI services, particularly when interacting with minors.

Key provisions of Georgia’s SB 540 include:

• Disclosure of AI Interaction: Operators must clearly and conspicuously inform minor account holders that they are interacting with a conversational AI service, not a natural person. This disclosure must appear at the beginning of each session and at least every three hours in continuous interaction.
• Prohibition of Harmful Content: Reasonable measures must be instituted to prevent the AI service from producing sexually explicit content, suggesting sexual conduct, or sexually objectifying minors.
• Prevention of Emotional Manipulation: Chatbots are prohibited from generating statements that would lead a reasonable person to believe they are interacting with a natural person, including simulating emotional dependence, romantic or sexual innuendos, or role-playing adult-minor romantic relationships.
• Parental Tools for Younger Minors: For users under 13, AI system operators must offer tools for parents or guardians to manage the minor’s privacy and account settings.
• Crisis Response Protocols: For all users, regardless of age, the bill requires the adoption of a protocol for chatbots to respond to user prompts regarding suicidal ideation or self-harm, including reasonable efforts to refer the user to crisis service providers.

The bill impacts all AI operators serving minor children and imposes a $10,000 fine from the Attorney General for violations. Notably, SB 540 does not include carve-outs for chatbots embedded within larger platforms, requiring major tech companies to comply.

California’s Executive Order on Generative AI Procurement

On March 30, 2026, California Governor Gavin Newsom signed Executive Order N-5-26, focusing on the responsible procurement and deployment of generative AI (GenAI) across state government. This order builds upon a prior executive order (N-12-23 from September 2023) and leverages the state’s significant purchasing power to influence market behavior and encourage responsible innovation.

Key directives of California’s Executive Order N-5-26 include:

• New Vendor Certification Requirements: The California Department of Technology (CDT) and Department of General Services (DGS) are directed to develop new procurement certifications. Companies seeking to contract with California agencies will need to attest to and explain their policies and safeguards regarding:
  • Exploitation or distribution of illegal content, such as child sexual abuse material and non-consensual intimate imagery.
  • Governance measures to reduce harmful bias in AI models.
  • Violations of civil rights and liberties, including free speech, voting, human autonomy, and protections against unlawful discrimination, detention, and surveillance.
• Review of Federal Supply Chain Risk Designations: The CDT’s Chief Information Security Officer (CISO) is tasked with independently reviewing federal supply chain risk designations for AI companies.
• Watermarking AI-Generated Content: Guidance will be issued for state departments and agencies to appropriately watermark AI-generated or significantly manipulated images or video, aligning with existing California law.
• Expanded Government Use of AI: The order facilitates state employee access to vetted GenAI tools with appropriate privacy and cybersecurity safeguards and identifies opportunities for GenAI to improve government services. This includes developing a pilot application or website to provide Californians with access to government services organized by life event.

This executive order is a strategic move by California to operate within the carve-outs identified by the federal framework for state government procurement and use of AI, showcasing a nuanced approach to AI Regulation amidst federal-state tensions.

The Delicate Balance: Federal Preemption vs. State Autonomy

The core tension in U.S. AI Regulation is the ongoing debate between establishing a unified federal framework and allowing states to develop their own targeted legislation. The White House explicitly argues that a fragmented patchwork of state AI laws imposes “undue burdens,” hindering innovation and national competitiveness. However, states like California are asserting their authority, particularly through their procurement power, to shape AI development and deployment. The federal framework does, importantly, preserve state authority in areas like child protection, fraud prevention, consumer protection, zoning, and state government’s own use of AI.

This divergence raises significant legal and operational questions around preemption, enforcement authority, and compliance burdens. The December 2025 Executive Order “Ensuring a National Policy Framework for Artificial Intelligence” by the Trump administration established an “AI Litigation Task Force” to challenge state AI laws deemed unconstitutional or preempted. However, the success of such broad preemption efforts remains uncertain, as Congress has previously declined to enact comprehensive federal preemption.

Challenges and the Path Forward for AI Regulation

Despite the comprehensive nature of the White House framework and the proactive state initiatives, significant challenges remain. One critical observation is the absence of a clear enforcement architecture within the federal framework. While the framework outlines goals and protections, it often lacks specified mechanisms to verify that AI platforms actually enforce these protections at the point of processing. This gap between policy intent and technical enforceability highlights a crucial area for future development in AI Regulation.

Moreover, the sheer volume of state legislative activity, with over 40 states introducing around 250 AI-related bills in 2025 alone, underscores the complexity. Companies operating across state lines face the daunting task of navigating a dynamic and potentially conflicting regulatory environment. This necessitates active tracking of regulatory developments, reassessment of AI footprints, and strengthened internal governance.

The evolving landscape of AI Regulation in the United States reflects a critical moment in technological governance. The federal government’s attempt to establish a unified national policy through its National Policy Framework for Artificial Intelligence seeks to prevent a regulatory “race to the bottom” or an unmanageable “patchwork” of state laws. Concurrently, states like Tennessee, Georgia, and California are demonstrating leadership by addressing specific, pressing concerns, from the ethical deployment of AI in mental health to the protection of minors online and the responsible use of generative AI within government operations. The path forward will undoubtedly involve continued negotiation, potential legal challenges, and a continuous adaptation of policies to keep pace with the rapid advancements in AI technology, all while striving to balance innovation with critical safeguards for society.