Texas moved faster than most states to regulate government AI use in 2025. Senate Bill 1964 took effect September 1, 2025, establishing procurement standards, disclosure requirements, and complaint mechanisms for state agencies deploying artificial intelligence systems. House Bill 149, the Texas Responsible Artificial Intelligence Governance Act (TRAIGA), took effect January 1, 2026, adding consumer protections, a regulatory sandbox, and Attorney General enforcement authority. Texas AI government transparency rules now shape how public sector vendors contract, document, and disclose automated systems.
This analysis summarizes bill scope and status, maps transparency publication duties, outlines vendor contract clauses procurement teams should expect, compares Texas with Colorado's approach, and links to AI regulation resources and AI chatbot deployments in government contexts.
Bill Scope and Status in 2026
Texas now operates under two complementary AI statutes: SB 1964 for government entity procurement and deployment, and HB 149 (TRAIGA) for broader consumer and developer obligations with state preemption of local AI ordinances. Both bills passed the 89th Legislature in 2025 with bipartisan support.
SB 1964 amends the Texas Government Code by adding Subchapter S on artificial intelligence systems. The Department of Information Resources (DIR) oversees minimum standards aligned with NIST's AI Risk Management Framework. State agencies must follow an AI system code of ethics covering transparency, fairness, accountability, and security. Heightened-scrutiny AI systems face additional minimum standards before deployment. The bill creates a Public Sector Artificial Intelligence Systems Advisory Board, an AI sandbox program for controlled testing, and enforcement mechanisms including contract-voiding provisions for noncompliance.
HB 149 adds Subtitle D to the Business and Commerce Code, establishing consumer protections, biometric data rules, civil penalties, and exclusive Attorney General enforcement. TRAIGA includes a regulatory sandbox enabling innovators to test AI systems without full licensing. The Attorney General must post a complaint mechanism on its website by September 1, 2026. Local governments cannot adopt conflicting AI ordinances under the preemption clause in Section 552.003.
| Statute | Effective date | Primary focus |
|---|---|---|
| SB 1964 | September 1, 2025 | Government AI procurement, ethics, disclosures |
| HB 149 (TRAIGA) | January 1, 2026 | Consumer protections, sandbox, AG enforcement |
| AG complaint portal | By September 1, 2026 | Public reporting channel for AI harms |
Transparency Publication Duties for Agencies
State agencies that procure or deploy public-facing AI systems must provide clear disclosure when citizens interact with automated systems and publish standardized notices describing system purpose, data sources, and privacy implications. Section 2054.707 of the Government Code requires agencies to disclose AI interactions to the public. Section 2054.711 mandates standardized notices for systems that make or inform consequential decisions.
Disclosure duties extend beyond chatbot greetings. Agencies must document which decisions are automated, which data sources feed models, and how individuals can contest or appeal outcomes. Impact assessments are required before deploying heightened-scrutiny systems. DIR provides guidance, training programs, and statewide oversight to ensure new AI technologies are vetted for fairness and security before production use.
SB 1964 also establishes a complaint web page on DIR's website where individuals can report AI systems that allegedly infringe constitutional rights, cause financial harm, or produce unlawful outcomes. Vendors selling to Texas agencies should anticipate audit requests for training data provenance, model versioning, and human oversight procedures tied to these disclosure obligations.
Vendor Contract Clauses to Expect
Texas government AI contracts increasingly require NIST AI RMF alignment, impact assessment cooperation, breach notification for model behavior changes, and termination rights when systems fail transparency or fairness standards. Procurement teams should prepare documentation packages before RFP responses.
- Ethics attestation: Vendor confirms compliance with the AI system code of ethics in Government Code Section 2054.702.
- Heightened-scrutiny standards: Systems affecting benefits, licensing, or law enforcement face additional minimum standards per Section 2054.703.
- Data management: Clauses governing training data, retention, and Texas public records obligations.
- Disclosure support: Vendor assists agency in drafting standardized public notices and interaction disclosures.
- Impact assessment: Vendor provides technical documentation for pre-deployment impact reviews.
- Enforcement remedies: Contract voiding and Comptroller referral for material noncompliance per Section 2054.709.
- Sandbox participation: Optional pathway for piloting systems under DIR oversight before full procurement.
TRAIGA adds developer-side obligations for systems sold to both government and commercial customers. Vendors should map which product features trigger biometric data rules, consumer disclosure requirements, and civil penalty exposure under Chapter 552. Government contracts may incorporate TRAIGA compliance by reference even when the primary statute is SB 1964.
Comparison with Colorado AI Rules
Texas emphasizes government transparency and Attorney General enforcement with state preemption, while Colorado's AI Act focuses on deployer obligations for high-risk automated decision-making technology (ADMT) in the private sector. The approaches complement rather than duplicate each other.
Colorado's amended AI Act (following SB 26-189) requires deployers of covered ADMT to implement risk management programs, conduct impact assessments, and notify consumers about consequential decisions. Texas SB 1964 applies similar impact assessment concepts specifically to state agencies rather than broadly regulating private deployers. TRAIGA's consumer protections overlap partially with Colorado's deployer duties but rely on AG enforcement instead of a private right of action.
Vendors serving both states should maintain separate compliance matrices. Colorado requires documented risk management for covered ADMT with specific effective dates for deployer obligations. Texas requires government-facing disclosure artifacts and standardized notices tied to procurement contracts. A chatbot sold to a Texas state agency triggers SB 1964 disclosure duties. The same product sold to a Colorado insurer may trigger ADMT deployer requirements under a different statutory framework.
| Dimension | Texas (SB 1964 / TRAIGA) | Colorado AI Act |
|---|---|---|
| Primary target | Government agencies + broad consumer rules | Private deployers of covered ADMT |
| Enforcement | DIR oversight, AG exclusive authority | AG enforcement, deployer duties |
| Local preemption | Yes (TRAIGA Section 552.003) | No equivalent broad preemption |
| Sandbox program | Yes (government and TRAIGA sandboxes) | Limited pilot concepts in amendments |
Frequently Asked Questions
When did Texas AI government transparency rules take effect?
SB 1964 took effect September 1, 2025. TRAIGA (HB 149) took effect January 1, 2026. The Attorney General must publish the TRAIGA complaint mechanism by September 1, 2026.
Which Texas agency oversees government AI standards?
The Department of Information Resources (DIR) provides guidance, training, minimum standards aligned with NIST's AI Risk Management Framework, and statewide oversight for state agency AI deployments.
Do Texas rules apply to local governments?
SB 1964 addresses state agencies and governmental entities. TRAIGA preempts local AI ordinances statewide, meaning cities and counties cannot adopt conflicting AI regulations. Local government AI use may still fall under SB 1964's broader governmental entity definitions depending on system scope.
What must vendors disclose in Texas government contracts?
Vendors should support agency disclosure duties including system purpose, data sources, privacy practices, and human oversight mechanisms. Heightened-scrutiny systems require additional documentation and impact assessment cooperation before deployment.
How does Texas differ from Colorado for AI vendors?
Texas centers on government procurement transparency and AG-enforced consumer rules with state preemption. Colorado focuses on private-sector deployer obligations for high-risk automated decision-making. Vendors serving both states need separate compliance documentation.
