Texas TRAIGA Scope Checker (HB 149)
Texas’s AI law (TRAIGA, HB 149) took effect January 1, 2026 — mostly prohibitions for everyone, disclosure duties mainly for government. Six questions sort which parts possibly reach your system.
Texas took the opposite road from Colorado and the EU. The Texas Responsible Artificial Intelligence Governance Act (TRAIGA, HB 149) — in force since January 1, 2026 — contains no high-risk register, no impact assessments and no general duty for private companies to disclose that they use AI. What it contains instead is a short list of things nobody may build AI to do, a set of operating rules for Texas government, and an enforcement model with a cure period and a narrow, carefully-worded reward for companies that find their own violations through a NIST-aligned internal review.
What this checks
The Texas Responsible Artificial Intelligence Governance Act (HB 149, effective January 1, 2026; Business & Commerce Code Title 11, Subtitle D, chapters 551–554) applies to a person who promotes, advertises or conducts business in Texas, produces a product or service used by Texas residents, or develops or deploys an AI system in Texas (§ 551.002) — there is no size threshold. Answer for one system at a time.
1 · Texas connection (§ 551.002)
2 · Are you a Texas governmental entity — or building this system for one?
“Governmental entity” covers state and local administrative units; hospital districts and institutions of higher education are carved out of the definition.
3 · Prohibited-purpose check (§§ 552.052–552.057) — these bind every person
Tick anything the system is designed or intended to do. “Intended” is the statute’s word: these are purpose tests, not outcome tests.
4 · Government-only duties
Answer honestly even if you are private — the result explains which of these can never reach you.
5 · Internal review
TRAIGA’s framework point is narrow — see the result for exactly what it does and does not do.
Texas Legislature Online — HB 149 enrolled text (89th Legislature, Regular Session)
Texas Legislature Online — HB 149 bill stages (signed June 22, 2025; effective January 1, 2026)
Key provisions: Bus. & Com. Code § 551.002 (applicability — no size threshold); § 552.051 (interaction disclosure: governmental agencies, plus healthcare providers); § 552.052 (manipulation); § 552.053 (social scoring — governmental entities); § 552.054 (biometric identification — governmental entities); § 552.055 (constitutional rights, “sole intent”); § 552.056 (unlawful discrimination, intent-based; disparate impact alone is not sufficient); § 552.057 (sexual content); §§ 552.101–552.106 (AG-exclusive enforcement, 60-day cure, penalties, and the § 552.105(e)(2)(D) liability bar for violations discovered through a compliant internal review). Retrieved November 5, 2026. The enrolled text controls — verify against the official text. Caution: “HB 149” of the 89th Legislature’s second called session is an unrelated bill; TRAIGA is 89(R) HB 149 only.
Quick answers
Does TRAIGA require my company to disclose that we use AI?
Generally no — if you are a private company. The § 552.051 disclosure duty for consumer-interacting AI systems falls on governmental agencies. The one private-sector reach is healthcare: a healthcare provider must disclose AI used in relation to health care service or treatment.
Private companies are still bound by TRAIGA’s prohibitions (manipulation, intentional unlawful discrimination, sole-intent rights infringement, sexual content) and by other Texas law the same act amended, including the CUBI biometric statute.
Is there a size threshold for TRAIGA?
No. Section 551.002 applies Subtitle D to any person who does business in Texas, produces a product or service used by Texas residents, or develops or deploys an AI system in Texas — with no employee, revenue or user-count floor.
That breadth is why the duties matter more than the scope test: most of the operative rules then narrow by actor (governmental entities) or by intent (the prohibitions).
Does following the NIST AI Risk Management Framework make us TRAIGA-compliant?
No — and the statute does not say that. Section 552.105(e)(2)(D) is a liability bar for violations a defendant discovers through an internal review process that substantially complies with the NIST AI RMF Generative AI Profile (or an equivalent recognised framework). It protects self-discovery and remediation; it is not an advance safe harbour or an exemption.
Separately, § 552.105(c) gives a defendant a rebuttable presumption that it used reasonable care as required by the chapter.
How to use
- Answer for one system at a time, and answer question 3 by the system’s designed purpose — TRAIGA’s prohibitions are intent tests, so “what is it for” is the legally load-bearing answer.
- Question 4 deliberately mixes duties: three of them can only ever bind governmental entities, one (healthcare) can bind a private provider. The result tells you which layer you are in.
- Read your tier and the notes under it, and check the key-dates table below. Copy the result link to share the assessment — it encodes only these answers.
About this tool
That shape produces two characteristic errors, which this checker is built to prevent. The first is importing other states’ duties into Texas: a private company generally owes no § 552.051 interaction disclosure (that duty sits on governmental agencies, with a single healthcare-provider exception), and the social-scoring and biometric-identification rules are likewise government-only. The second error is assuming “no duties” means “no law”: the prohibitions on manipulation, intentional unlawful discrimination, sole-intent rights infringement and sexual content bind every person with a Texas connection, and the same act tightened the CUBI biometric statute and the Texas privacy act.
Scope itself is unusually broad: § 551.002 has no employee, revenue or user threshold at all — doing business in Texas, having Texas-resident users, or developing or deploying AI in Texas is enough. The Attorney General enforces exclusively (there is no private right of action), must offer 60 days to cure, and penalties scale from $10,000–$12,000 per curable violation to $80,000–$200,000 per uncurable one, plus $2,000–$40,000 per day while a violation continues. The act also creates a regulatory sandbox (chapter 553, up to 36 months, with the prohibitions non-waivable) and a Texas Artificial Intelligence Council that can advise but cannot bind (chapter 554).
How it works
TRAIGA’s architecture, in the order this checker applies it. Layer one is reach: § 551.002 applies Subtitle D to a person who (1) promotes, advertises, or conducts business in Texas; (2) produces a product or service used by Texas residents; or (3) develops or deploys an AI system in Texas. “AI system” (§ 551.001) is any machine-based system that infers from its inputs how to generate outputs — content, decisions, predictions or recommendations — that can influence physical or virtual environments. “Consumer” means a Texas resident acting in an individual or household context, not a commercial or employment one.
Layer two is the prohibitions (Subchapter B), which bind every person reached: § 552.052 (developing or deploying AI that intentionally aims to incite self-harm, harm to others or criminal activity); § 552.055 (AI developed or deployed with the sole intent to infringe constitutional rights); § 552.056 (AI developed or deployed with the intent to unlawfully discriminate — disparate impact alone cannot prove that intent); § 552.057 (AI whose sole intent is sexual material violating the Penal Code, and AI text conversations that simulate sexual conduct while impersonating a child). Layer three is the government-only rules: § 552.051(b) interaction disclosure by governmental agencies (plain language, clear and conspicuous, no dark patterns), § 552.053 social scoring, § 552.054 biometric unique identification — plus § 552.051(f), the healthcare-provider disclosure, which is the single affirmative duty that reaches the private sector.
Layer four is enforcement and its incentives (§§ 552.101–552.106): AG-exclusive, no private right of action, civil investigative demands, a 60-day cure window, penalties of $10,000–$12,000 (curable), $80,000–$200,000 (uncurable) and $2,000–$40,000 per day (continued), no penalty for a system that has not been deployed, and agency sanctions after an AG-recommended finding. The framework incentives sit inside § 552.105: a rebuttable presumption of reasonable care (subsection (c)), and the subsection (e)(2)(D) liability bar for violations discovered through an internal review substantially complying with the NIST AI RMF Generative AI Profile or an equivalent framework. Both reward documented self-scrutiny; neither is an exemption.
Worked examples
Private SaaS chatbot for Texas customers: A private company sells a support chatbot used by Texas residents; it is not designed to do any prohibited thing. Nexus: yes. Government: no. Result pattern: likely out of the affirmative duties — no TRAIGA disclosure duty for the chatbot — with the prohibitions and other Texas law still applying.
City benefits-eligibility assistant: A Texas city deploys an assistant that chats with residents and scores their likely benefit eligibility. Nexus: yes. Government: yes, with interaction and scoring both ticked. Result pattern: likely in scope of the governmental duties — § 552.051 disclosure at interaction, and § 552.053’s social-scoring limits to work through with counsel.
Clinic triage messaging tool: A private Texas clinic uses AI to message patients about their care. Government: no — but the healthcare box in question 4 is ticked. Result pattern: possibly in scope, for the § 552.051(f) healthcare disclosure specifically.
Key dates & data passport
| Date | Milestone |
|---|---|
| June 22, 2025 | HB 149 (TRAIGA) signed by the Governor |
| January 1, 2026 | TRAIGA takes effect (Act § 10) — chapters 551–554 in force |
| September 1, 2026 | Deadline for the Attorney General to post the § 552.102 complaint information and mechanism (Act § 8) |
Data passport
| Legal source | Texas Legislature Online — HB 149 (89th Legislature, Regular Session), enrolled text, capitol.texas.gov; Business & Commerce Code Title 11, Subtitle D, chapters 551–554 |
|---|---|
| Applies from | January 1, 2026 (Act § 10) |
| Last verified | November 5, 2026 — enrolled text and official bill stages read directly |
| Watch owner | verystep-regulatory-watch (monthly check against the official source) |
Frequently asked questions
When did TRAIGA take effect?
January 1, 2026 (Act § 10). It was signed on June 22, 2025. The Attorney General was separately required to post the § 552.102 complaint mechanism by September 1, 2026.
Can a consumer sue us under TRAIGA?
No. Enforcement is exclusive to the Attorney General (§ 552.101), who must give a 60-day notice-and-cure opportunity before bringing an action (§ 552.104). Other laws — civil-rights, consumer-protection, biometric-privacy — have their own causes of action and are unaffected.
Our AI sometimes produces biased outcomes. Is that a § 552.056 violation?
Not by itself. Section 552.056 prohibits developing or deploying AI “with the intent to unlawfully discriminate”, and § 552.056(c) states expressly that disparate impact alone is not sufficient to demonstrate intent. Outcomes can still violate other anti-discrimination laws — the intent point is specific to this section.
What is the AI regulatory sandbox?
Chapter 553 lets the Department of Information Resources approve participants to test AI systems for up to 36 months with quarterly reporting. The Chapter 552 prohibitions cannot be waived inside the sandbox (§ 553.051(e)).
Does TRAIGA apply to our internal-only tool?
The § 551.002 connection test can still be met (you develop or deploy AI in Texas), so the prohibitions apply to it. Most affirmative duties, however, are aimed at governmental entities and consumer-facing systems, so an internal private tool usually lands in this checker’s “prohibitions still apply” tier.
Is this legal advice?
No. It is a plain-language walk through the enrolled text. Intent questions and agency status are fact-specific — have Texas counsel review before relying on it.
First time embedding? Read the step-by-step guide →
More free tools
EU AI Act Article 50 Checker
Four Article 50 duty blocks — interaction disclosure, machine-readable marking, emotion-recognition notice, deep fake and public-interest text disclosure — sorted by whether you are the provider, the deployer, or both.
California SB 942 Scope Checker
Five questions tell you whether California’s AI Transparency Act (SB 942, after SB 1000 removed the million-user threshold) likely makes you a “covered provider” with latent-disclosure and verification-tool duties.
California ADMT Scope Checker
Three gates — CCPA business thresholds, significant decisions, human review — tell you whether California’s automated decisionmaking rules likely cover your tool.
Illinois AI in Employment Checker (HB 3773)
From January 1, 2026, Illinois employers must give notice when AI touches hiring and other listed employment decisions — and may not use AI with a discriminatory effect, or ZIP codes as a proxy. Four questions, one narrow law.
Colorado AI Act Scope Checker (SB 26-189)
Colorado repealed its first AI Act and replaced it with an ADMT notice regime — SB 26-189. Five questions show whether your automated decision-making technology is possibly a “covered ADMT” from January 1, 2027.
Invoice Generator
Create professional invoices in seconds — download as PDF, no sign-up.