Disruption

Texas lawmaker unveils sweeping AI bill for 2025

Rep. Giovanni Capriglione’s (R) legislation could serve as a model for red states.
Texas state Rep. Giovanni Capriglione (R) is taking the lead on artificial intelligence legislation. (Courtesy of Vote Giovanni)

A Texas GOP lawmaker has released a highly anticipated draft of a comprehensive artificial intelligence regulation bill that could become a blueprint for how red states regulate the fast-moving technology.

Rep. Giovanni Capriglione (R), chair of the Innovation and Technology Caucus, unveiled the legislation Monday, culminating a months-long bill drafting process that he said included input from more than 300 industry experts and others. He intends to formally introduce the legislation for the 2025 legislative session.

“Our work has sought to adapt best practices into a uniquely Texan framework that balances Texas’ commitment to fostering innovation while safeguarding fundamental rights and ensuring ethical AI deployment,” Capriglione said in an email to stakeholders, which invited their feedback.

Capriglione described the 41-page draft legislation, dubbed the Texas Responsible AI Governance Act, as taking a risk-based approach to AI regulation. Specifically, it seeks to guard against algorithmic discrimination by automated decision-making systems.

The bill exempts AI models that are used for research or testing as part of a “sandbox program,” and it does not cover low-risk AI systems. Capriglione said that would help ensure “Texas remains a friendly environment for innovation.”

Capriglione has emerged as a key leader on AI in the Texas House. He led the passage of legislation last year that established the governor’s AI advisory council to study government’s AI use. Capriglione now serves on that council and chairs a select committee on AI. He is also vice chair of the National Conference of State Legislatures’s AI task force and is on the steering committee of a bipartisan multistate working group on AI.

State lawmakers throughout the country have taken the lead on regulating AI because Congress has so far not acted. Much of the work to date has focused on the danger of election-related and intimate deepfakes.

Read more: AI bills poised to boom again in ’25

Lawmakers are also working on a comprehensive, risk-based regulation that targets AI systems used to help make consequential decisions about people’s lives in areas such as housing, employment and lending. The concern is that without testing and regulation, those systems’ algorithms could discriminate against people based on race, gender or other protected classes.

Connecticut Sen. James Maroney (D), who launched the multistate working group, released a sweeping risk-based AI regulation bill this year that is widely viewed as a blue state blueprint. Maroney’s bill ran into opposition from Gov. Ned Lamont (D) and was not passed, but a version of the legislation was enacted in Colorado.

Lawmakers in a dozen or more states are expected to introduce similar legislation in 2025, according to members of the working group. If that happens, it will represent a significant acceleration of state-level efforts to place comprehensive guardrails on AI-powered technologies. Other lawmakers plan to take a product liability approach to AI regulation.

Capriglione’s bill is noteworthy because it has the potential to emerge as a red state analog to Maroney’s legislation, much as Republican-led states have passed lighter touch consumer data privacy laws. Capriglione, who has a background in business and technology, has argued that Texas can be a “global center of AI” and should lead in the development of a “responsible, pro-innovation framework.”

Throughout the bill drafting process, Capriglione highlighted the potential benefits of AI and its dangers, such as the erosion of human autonomy and the loss of truth. He has argued that regulations will bring certainty to investors while protecting individuals’ privacy and rights.

In a PowerPoint presentation this year, Capriglione wrote that “AI threats are too serious to piecemeal a solution” and said lawmakers should not repeat the mistake of waiting 30 years to regulate consumer data privacy. In 2023, Capriglione led the passage of a comprehensive consumer data privacy law.

On a slide titled: “Why must Texas Act?” Capriglione listed federal inaction and the opportunity to “accelerate responsible AI.”

Under Capriglione’s proposed legislation, developers and deployers of high-risk AI systems would have to take “reasonable care” to ensure that the systems do not discriminate. This would include regular testing and a requirement that AI systems found to be out of compliance with the law must be disabled or recalled until the problem can be resolved.

Another provision of the bill would require that consumers be informed if they are interacting with a high-risk AI system and told how it might be used to make a consequential decision about them. Consumers would also have the right to appeal a consequential AI-driven decision.

Other elements of the bill would prohibit using AI to manipulate human behavior, classify people based on “social scoring,” or collect biometric data from users, among other prohibitions.

The proposed law would be enforced by the state’s attorney general with fines of up to $100,000 for certain violations. AI developers and deployers would have 30 days to resolve any compliance issues before enforcement action was taken.

To support AI innovation in Texas, the bill would also create an AI workforce development grant program to “prepare Texas workers and students for employment in the rapidly growing artificial intelligence industry.”

If passed, the bill would take effect Sept. 1, 2025.