How Other States Are Approaching AI

California SB 53 and Texas HB 149

Florida’s choices are easier to evaluate when compared with other states. California and Texas offer useful examples of different approaches, although neither is a perfect template for Florida.

California: Transparency and safety for frontier AI

California enacted SB 53, the Transparency in Frontier Artificial Intelligence Act, on September 29, 2025. The law establishes requirements for covered frontier AI developers, including publication of a framework describing their approach to safety and risk, specified reporting mechanisms for critical safety incidents, and protections for certain employees who disclose covered risks or violations. It also provides for enforcement and creates a public-computing initiative known as CalCompute.

SB 53 should be distinguished from California’s earlier SB 1047 proposal, which Governor Gavin Newsom vetoed in September 2024. SB 1047 and SB 53 are separate measures with different provisions. Treating the veto of the earlier bill as if California rejected all frontier-AI regulation would obscure the subsequent legislative action. California’s approach is relevant to Florida because it illustrates a model focused on certain advanced AI developers and transparency about safety practices, rather than a single rule governing every use of AI.

For Florida policymakers, the questions include whether comparable disclosure requirements would be useful, how to define covered developers, how to protect legitimate confidential information, and what enforcement mechanism would be appropriate.

Texas: Targeted statutory prohibitions and enforcement

Texas enacted HB 149, the Texas Responsible Artificial Intelligence Governance Act, in 2025, with an effective date of January 1, 2026. The measure is Chapter 1174 of the Acts of the 89th Legislature, Regular Session, and its principal AI provisions are codified in Chapter 552 of the Texas Business & Commerce Code. The law uses targeted prohibitions and specified obligations rather than treating every AI system or use identically. It also assigns enforcement authority to the Texas Attorney General for the provisions in Chapter 552 and does not create a private right of action under that chapter. The details matter. Texas’s law should not be described as a blanket prohibition on high-risk AI or as a universal licensing regime. Its coverage depends on the statutory definitions, conduct addressed and applicable exceptions.

For Florida, the Texas model raises questions about whether legislation should focus on prohibited conduct, establish disclosures for specified uses, rely primarily on government enforcement, or create additional remedies. Those choices affect the compliance burden on businesses as well as the ability of individuals and public agencies to respond to harm.

The comparison for Florida

California and Texas demonstrate that state AI legislation can take materially different forms. California’s SB 53 emphasizes transparency and safety obligations for covered frontier developers. Texas’s HB 149 establishes targeted rules and enforcement provisions for specified AI-related conduct.

Florida need not choose one state’s approach in its entirety. It could adopt rules for particular applications, focus on consumer and child protections, establish requirements for government use, address frontier-model risks, or prioritize the infrastructure supporting AI. Those policy choices could be combined, sequenced or left to existing law where lawmakers conclude that new requirements are unnecessary. The important question is not simply whether Florida should regulate AI. It is what problem a proposed rule is intended to solve, which entities and conduct it covers, how it interacts with existing law, and whether its expected benefits justify its costs.