Texas Just Passed One of the Biggest AI Laws in America

New Rules Take Effect January 2026

Texas is now one of just four states with a comprehensive law governing artificial intelligence.

Governor Greg Abbott signed the Texas Responsible Artificial Intelligence Governance Act on June 22, 2025, creating new rules for how AI can be developed, tested, and used across the state.

The law bans certain harmful uses of AI, gives developers a three-year testing window, and puts the Attorney General in charge of enforcement.

But what Texas built, and what it left out, tells you a lot about where AI regulation is heading in America.

Fourth State to Regulate AI

Texas joins Colorado, Utah, and California as states with broad AI legislation. House Bill 149 passed the state House with a 146-3 vote before landing on the governor’s desk.

Representative Giovanni Capriglione of Southlake authored the bill, which he called one of the biggest AI frameworks any state will pass.

The law covers anyone doing business in Texas or producing AI products consumed by Texas residents, which means companies far outside state lines will need to pay attention.

Original Bill Got Gutted

The law that passed looks nothing like the one Capriglione first filed in December 2024. That version ran 43 pages and took aim at high-risk AI systems with strict requirements for developers.

Industry groups pushed back hard, warning the rules would crush innovation.

By March 2025, lawmakers had stripped out most of the aggressive provisions and refiled a slimmer 31-page version.

The final product focuses mainly on government AI use and specific harmful applications rather than broad industry oversight.

What the Law Bans

The legislation draws hard lines around certain AI uses.

Developers cannot build systems designed to discriminate against protected classes, manipulate human behavior, create child exploitation material, or generate nonconsensual deepfakes.

AI systems intended to encourage self-harm, violence, or criminal activity are also prohibited.

The law makes clear that existing constitutional protections apply to AI, meaning systems cannot be built with the sole intent of violating individual rights.

Three Years to Experiment

Texas created a regulatory sandbox that lets companies test AI systems for up to 36 months without full compliance requirements.

Developers must apply through the Department of Information Resources and submit a benefit assessment covering impacts on consumers, privacy, and public safety.

Once approved, participants file quarterly reports detailing performance metrics, risk mitigation efforts, and user feedback. The sandbox is designed to encourage innovation while keeping some guardrails in place.

Advisory Council Has No Teeth

The law creates a seven-member Texas Artificial Intelligence Council appointed by the governor, lieutenant governor, and speaker of the house.

Members must be Texas residents with expertise in areas like AI, ethics, law enforcement, or constitutional rights. The council will train government workers on AI and issue reports to the legislature.

But here is the catch: it cannot create any binding rules or regulations. It advises, it does not command.

Attorney General Runs Enforcement

All enforcement power sits with the Texas Attorney General.

Violators face civil fines up to $200,000 for serious offenses that cannot be fixed and $40,000 per day for ongoing violations.

The law includes a 60-day cure period, giving companies a chance to correct problems before penalties kick in. There is no private right of action, meaning individual Texans cannot sue companies directly.

They can file complaints through an online portal, but the AG decides whether to act.

Government Gets Stricter Rules

State agencies face the toughest requirements under the new law.

Any government website using AI to interact with citizens must disclose that fact through clear, plain language.

Healthcare providers must tell patients when AI is involved in their treatment, either before service begins or as soon as possible in emergencies.

The law puts heavier scrutiny on public sector AI use to ensure systems uphold constitutional rights.

Taxpayers Foot the Bill

The framework costs about $25 million through 2027, according to the Legislative Budget Board. That money funds 20 new state employees, including 12 in the Attorney General’s office.

The Department of Information Resources needs around $3 million annually to run the sandbox program, while the AG’s office requires $4 million per year to hire outside AI and machine learning experts.

After the initial setup, ongoing costs run about $10 million every year.

Federal Moratorium Failed

Congress nearly wiped out state AI laws entirely. The House passed a budget bill in May 2025 that included a 10-year moratorium on state and local AI regulation. Forty state attorneys general sent a letter opposing it.

Republican and Democratic lawmakers both criticized the overreach. When the bill reached the Senate in July, lawmakers stripped the AI moratorium by a 99-1 vote.

President Trump signed the budget bill on July 4, 2025, without the provision. State AI laws survived.

Texas Beats Colorado by One Month

The Texas law takes effect January 1, 2026, one month before Colorado’s AI Act goes live on February 1. That timing makes Texas an early test case for how comprehensive state AI regulation actually works in practice.

Businesses operating in both states will need to navigate two different frameworks almost simultaneously. Representative Capriglione has said Texas aims to be a model other states can follow.

Clock is Ticking for Businesses

Companies have until New Year’s Day 2026 to prepare.

That means reviewing any AI systems they develop or deploy in Texas, checking whether those systems fall under the prohibited uses, and deciding whether to apply for the regulatory sandbox.

Texas is also building a new AI division within the Department of Information Resources to help state and local governments adopt AI tools.

The sandbox applications, the council appointments, and the enforcement mechanisms will all be taking shape over the coming months.

This article was created with AI assistance and human editing.

Read more from this brand:

• Democrats Want to Stop AI From Discriminating Against You
• U.S. Town Famous for Banning WiFi Finally Surrenders to the Internet
• New York Will Force Every Store to Accept Cash by March 2026