BlogLine

A new sheriff in town?: Texas legislature passes the Texas Responsible AI Governance Act

6/11/25

By: Danielle A. Ocampo

There’s a new sheriff in town when it comes to regulating the wild west of AI Governance. The Texas legislature passed a new comprehensive AI law, H.B. 149, the Texas Responsible AI Governance Act (TRAIGA), that will arguably be the toughest in the nation.   

TRAIGA takes a broader approach to AI regulation than its predecessor, the Colorado AI Act (CAIA). Similar to the European Union AI Act, TRAIGA prohibits certain uses of AI that are deemed to have unacceptable risks without consent: 

TRAIGA requires government agencies to disclose when Texans are interacting with an AI system on a state agency website.  

The Texas Attorney General’s office serves as the sheriff itself with full enforcement capabilities, including the authority to fine violators up to $200,000 per continued daily non-compliance after failing to cure within a 60-day notice period. Developers and deployers have affirmative defenses at their disposal in addition to curing, namely, adversarial testing or red-teaming of AI systems, compliance with the latest NIST Risk Management Framework or ISO/IEC 42001 and appropriately following state agency guidelines. 

TRAIGA also creates a ten-member Texas Artificial Intelligence Council charged with issuing reports and recommendations on the ethical, compliant and safe use and development of AI.  

To encourage innovation that is balanced with regulation, TRAIGA creates an AI Regulatory Sandbox Program to allow controlled environments where program participant-developers can test and deploy AI systems for up to 36 months without penalties.  

Finally, though consumers are not afforded a private right of action, they have to a right to obtain “clear and meaningful explanations” from the developer of the AI system’s role in the decision-making process and the main elements of the decision taken to appeal AI decisions that have an adverse impact on health, welfare, safety or fundamental rights. Because the law defines AI systems to include machine learning and related technologies, requiring disclosure of the AI System’s role in a decision-making process is distinct from requiring transparency for one of the most difficult tasks with AI governance: being able to explain an output generated from a machine learning model’s “black box.” Previous iterations of the bill mandated certain AI governance documentation. Despite the removal of these provisions from the current TRAIGA text, best practices for any AI governance program require documentation regarding the design, development and implementation of AI, especially to respond to consumer appeals under TRAIGA.   

If signed by Texas Gov. Greg Abbott by June 22, 2025, the TRAIGA will be effective January 1, 2026. The town buzz is whether TRAIGA really is the newest sheriff in town. With the current federal administration’s “One Big Beautiful Bill” receiving momentum, the federal government would be the newest sheriff preempting all state AI enforcement efforts, including TRAIGA.  

For more information, please contact Danielle A. Ocampo at danielle.ocampo@fmglaw.com or your local FMG attorney.

Information conveyed herein should not be construed as legal advice or represent any specific or binding policy or procedure of any organization. Information provided is for educational purposes only. These materials are written in a general format and are not intended to be advice applicable to any specific circumstance. Legal opinions may vary when based on subtle factual distinctions. All rights reserved. No part of this presentation may be reproduced, published or posted without the written permission of Freeman Mathis & Gary, LLP.