'Big Beautiful' Reconciliation Bill Will Upend States’ Bottom-Up AI Governance

The House’s “Big, Beautiful Tax Bill,” also known as the reconciliation bill, proposes to ban state regulations on artificial intelligence (AI) for up to ten years. The legislation's language emphasizes a moratorium on the enforcement of existing state laws regulating AI models and automated decision systems, as well as future state-driven regulations. Without a federal governance framework currently in place and implemented, the provision stands to override at least a year’s worth of passed AI policy legislation at the state level and potentially inhibit experimentation and innovation around the use of the emerging technology.

Proponents argue that AI is best regulated at the federal level because a patchwork of state laws and regulations can overburden companies operating in multiple regions of the United States. State laws could also complicate federal initiatives around AI, which is another component of the bill. In December 2024, the House’s Task Force on Artificial Intelligence report was an example of what a federal policy framework could look like.

Proponents’ claims have some merit, as national consistency around AI policy would help refine approaches to risk mitigation while maximizing AI’s utility across diverse and complex regional economies and labor markets. However, as currently configured, this ban would simply step on a state’s ability to respond more timely to specific community needs, and create a regulatory vacuum, given that no comprehensive federal AI governance legislation is on the books.

Since ChatGPT ignited an AI revolution, states have taken the lead in experimentation and innovation in AI policy. According to the National Conference of State Legislatures, last year alone, at least 45 states introduced AI bills, and 31 states, plus Puerto Rico and the Virgin Islands, passed legislation or resolutions. Spanning the right-left continuum from Florida to Hawaii, state legislatures have been able to agilely meet the needs of their citizens and businesses. Legislation has addressed elections, child protection, education, use of AI in government, and private sector use.

The provision has already seen substantial pushback. Senate Democrats are expected to challenge this possible moratorium under the pretext of the Byrd Rule, which prohibits the inclusion of provisions in reconciliation bills that are “extraneous” to the federal budget.

Congressional leaders will make tough decisions this summer, but dropping the moratorium on state AI laws should be a no-brainer.

Senate Republicans have pushed back right out of the gate, including Missouri Senator Josh Hawley and Senator Marsha Blackburn, who took issue with how the bill would steamroll the ELVIS Act in her state, which protects artists from AI-generated deep fakes. Forty bipartisan state attorneys general from Arkansas to Arizona to Michigan have voiced strong opposition.

State-level innovation has played a critical role in advancing AI education, workforce development, policy experimentation, and access to emerging technologies. Benefits include the opportunity to act faster to test and refine programs, address specific needs of communities, create case studies that can inform national guidance, and address the needs of targeted audiences such as those in rural and low-income communities. Examples in education and workforce development include:

• Utah is one of the states leading the adoption of AI in K–12 education. They pioneered a regulatory sandbox to test policy innovations that have inspired other states and Canada to do the same. In April 2024, the state board of education established an AI framework to help guide localities in the responsible use of AI in P–12 education. They also launched one of the country’s first statewide funding solicitations for AI tools in schools, and invested in training and professional development that brought together more than 700 educators across the state. In addition, the Utah Office of Artificial Intelligence Policy (OAIP) is implementing a new framework to evaluate AI tools, promoting responsible and transparent innovation.
• The Massachusetts Department of Elementary and Secondary Education launched its multi-year strategy to integrate AI into its K–12 education system. This initiative aims to responsibly incorporate AI into classrooms, focusing on educator training, AI literacy, and policy development over the next three years.
• The California Department of Education has developed resources to integrate AI skills and computer science standards into K–12 education. This integration focuses on fostering computational thinking, problem-solving, and innovation, preparing students to engage with emerging technologies across various industries. The curriculum emphasizes interdisciplinary learning, connecting AI concepts with subjects like healthcare, engineering, agriculture, and art. At the postsecondary level, Governor Gavin Newsom and NVIDIA Founder and CEO Jensen Huang inked a major deal to scale AI education across the state’s community colleges, supporting initiatives around curriculum and certifications, procuring equipment and software, creating AI labs and workshops, and more.

Congressional leaders will make tough decisions this summer, but dropping the moratorium on state AI laws should be a no-brainer. Instead, Congress should continue to learn from state ingenuity and innovation as it formulates the best way to establish a national policy and regulatory framework relating to AI.