Skip to content

Texas TRAIGA (HB 149) — Responsible AI Governance Act

House Bill 149 — Texas Responsible Artificial Intelligence Governance Act (TRAIGA) was signed by Governor Greg Abbott on 2025-06-22 and enters into force on 2026-01-01.

Significance:

  • The third US comprehensive state AI law (after the Colorado AI Act and California SB 53).
  • But structurally clearly different from Colorado / California — TRAIGA focuses on the explicit prohibition of harmful uses, rather than layered obligations on a generic “high-risk AI system”.
  • It exhibits features of the Republican version of US AI legislation: prohibition > compliance > disclosure.

“Artificial intelligence system” means any machine-based system that, for any explicit or implicit objective, infers from the inputs the system receives how to generate outputs, including content, decisions, predictions or recommendations, that can influence physical or virtual environments.

→ Close to the OECD / EU AI Act definition; broad coverage.

Core prohibitions (applicable to all entities)

Section titled “Core prohibitions (applicable to all entities)”

TRAIGA expressly prohibits the development or deployment of AI systems in the following categories:

  • AI systems that, by intent, induce or encourage self-harm or criminal activity.
  • Corresponds to EU AI Act Article 5(1)(a).
  • AI systems that generate child sexual abuse imagery (CSAI).
  • AI systems that generate deep fake pornography.
  • Text-conversation AI that produces simulative / descriptive content which also imitates or impersonates a child.

These prohibitions are more concrete than federal COPPA or comparable California laws.

  • Government entities are prohibited from using AI systems for “social scoring”.
  • Does not apply to the private sector.
  • Corresponds to EU AI Act Article 5(1)(c).

Additional obligations for government entities

Section titled “Additional obligations for government entities”

1. Mandatory disclosure of AI interactions

Section titled “1. Mandatory disclosure of AI interactions”
  • Consumers (citizens) must be clearly informed that they are “interacting with an AI” before or at the point of interaction with a government AI system.
  • This is stricter than the Colorado AI Act’s consumer-notice requirement (TRAIGA explicitly says “before or at the point”).

See above.

  • Government is prohibited from using AI to independently identify individuals based on biometric data obtained from publicly available sources (web, surveillance, etc.) unless the individual has consented.
  • Precisely targets Clearview-AI-style law-enforcement AI.
  • The Texas Attorney General.
  • 60-day cure period — notice must be given before suit.
  • Citizens may not directly sue entities for TRAIGA violations.
  • TRAIGA expressly preempts AI ordinances of cities and counties within the state.
  • Prevents Austin, Dallas, Houston and others from legislating on their own.
  • Similar logic to EO 14365 (here the state preempts localities, rather than the federal government preempting states).
Violation typePenalty (per violation)
Curable violation (not cured / breach of cure statement)$10,000 – $12,000
Uncurable violation$80,000 – $200,000
Continuing violation (beyond the cure period / without a cure statement)$2,000 – $40,000 / day

Comparison with the Colorado AI Act / California SB 53

Section titled “Comparison with the Colorado AI Act / California SB 53”
DimensionColorado AI ActCalifornia SB 53Texas TRAIGA
PositioningAnti-discrimination / high-risk AIFrontier AI transparencyProhibition of specific harmful uses
ScopeDeveloper + DeployerLarge frontier developers ≥ $500MAll entities (some provisions government-only)
Compliance burdenHigh (risk management + impact assessment + notice)Medium (framework disclosure + incident reporting)Low (as long as you avoid the prohibited acts)
PenaltiesPer CPA$1M / violationUp to $200K / violation + per-day accrual for continuing violations
Local preemptionUnclearNoneExplicitly preempts cities and counties
Party leanDemocratic Party leadDemocratic Party leadRepublican Party lead

TRAIGA’s Republican character:

  • Prohibition of harmful uses (clear red lines) rather than procedural compliance (documentation burden).
  • Preserves commercial freedom.
  • A clear federal / state / local hierarchy (state preempts localities).

Because of its concise prohibition list, TRAIGA is comparatively less vulnerable to preemption by EO 14365 (since what it regulates is a broadly shared notion of harmful use). However, if the DOJ AI Litigation Task Force challenges the government AI-interaction disclosure clause, a federal / state conflict could be triggered.

SourceLink
Full text (state legislature PDF)capitol.texas.gov/…/HB00149I.pdf
Bill history (legislature)capitol.texas.gov/…/HB149
LegiScan full textlegiscan.com/TX/text/HB149
Norton Rose Fulbright commentarynortonrosefulbright.com/…/traiga
K&L Gates commentaryklgates.com/…/TRAIGA-signed
IAPP Newsiapp.org/news/a/governor-signs-texas-traiga
DateEvent
2025-06-22Signed by Governor Abbott
2026-01-01In force