The Attorney General is the plaintiff class, not a private bar. Texas HB 149, the Responsible Artificial Intelligence Governance Act, took effect 1 January 2026. The AG website mechanism is due 1 September 2026.

The news

The Texas Legislature enrolled-bill summary for HB 149, House Author Capriglione et al., Senate Sponsor Schwertner et al., titles the measure the Texas Responsible Artificial Intelligence Governance Act. It amends the Business and Commerce Code and Government Code to establish a regulatory framework for the use of AI systems in Texas.

Among other provisions, the summary says the bill: establishes mandatory disclosure requirements for certain users of AI systems, including governmental agencies and health care service providers; prohibits the development or deployment of an AI system to incite or encourage certain activities, gather certain biometric data in violation of an individual's rights, otherwise impair a person's constitutional rights, or unlawfully discriminate against a protected class; prohibits the use or deployment of an AI system to create a social scoring system; prohibits the development or distribution of an AI system to produce, assist or aid in producing, or distribute certain sexually explicit content or child pornography; creates a civil penalty and grants the attorney general authority to investigate, send notice, and bring an action against a person who violates the bill's provisions; and provides for a regulatory sandbox program enabling a person to obtain legal protection and limited market access to test AI systems, with an annual report.

The summary also creates the Texas Artificial Intelligence Council to analyze opportunities for improvement, conduct training programs for state agencies and local governments, make recommendations, and issue reports to the legislature.

The Senate amendments section-by-section analysis in the legislative PDF states SECTION 8: not later than 1 September 2026, the attorney general shall post on the attorney general's Internet website the information and online mechanism required by the added Business and Commerce Code section, and SECTION 9: this Act takes effect 1 January 2026.

Who is bound

The Texas Legislature shipped HB 149. Bound are persons who develop or deploy AI systems in the ways the chapter prohibits or who owe the disclosure duties (named classes include governmental agencies and health-care service providers). The Attorney General investigates and sues. The Council is a state body, not a licensing board. Sandbox applicants are a separate, opt-in class.

What's new

A Texas-wide prohibited-use and disclosure statute with AG enforcement and a sandbox, effective 1 January 2026. It is not Colorado's ADMT notice regime and not California's frontier-framework statute. Secondary trackers describe TRAIGA as intent-based. That characterization is not in the enrolled summary quoted here. Mens rea is UNKNOWN as a statutory test until counsel reads the enrolled text.

What it does not settle

Date: 1 January 2026 effective; AG website mechanism due 1 September 2026. Region: Texas. Price: civil penalty exists per the summary; dollar amounts are in the enrolled text, not copied here. Hardware: none. Data: biometric gathering is among the prohibited topics as summarized. License: not a software license. Sandbox legal protection is limited-market and time-bounded per the summary; duration UNKNOWN from the summary page.

What to do now

Texas health systems and state agencies using chatbots or scoring tools should locate the disclosure duty in the enrolled chapter and the AG's forthcoming web mechanism. Product counsel for consumer AI used in Texas should map the prohibited-use list (social scoring, CSAM, biometric collection in violation of rights, unlawful discrimination) against actual product behavior.

The dollar civil penalty is UNKNOWN without the enrolled chapter line cite.