Texas has a state AI law on the books. The Texas Responsible Artificial Intelligence Governance Act (TRAIGA) is House Bill 149 from the 89th Legislature. It took effect January 1, 2026 (enrolled bill summary; enrolled text; Section 10).
This post is a plain reading of the enrolled statute and the Texas Attorney General’s public overview for people who run systems in Texas. It is not a sales deck. If you already inventory your agents and decision tools, use this as a cross-check against the statute. If you do not, start with the checklist at the end.
Who it covers
TRAIGA 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 artificial intelligence system in Texas (Bus. & Com. Code § 551.002).
An “artificial intelligence system” under the Act is a machine-based system that, for an 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 (Bus. & Com. Code § 551.001(1); Texas AG overview).
A “consumer” under the Act is a Texas resident acting only in an individual or household context, not in a commercial or employment context (Bus. & Com. Code § 551.001(2)).
If your stack uses models, agents, or decision tools that fit that definition and you operate in Texas, Bus. & Com. Code ch. 551 coverage applies. Chapter 552 also preempts city and county AI ordinances, so you deal with one statewide set of rules rather than a local patchwork (Bus. & Com. Code § 552.003).
What the Act requires you to notice first
Disclosure when government or health care AI talks to a consumer.
A governmental agency that makes an AI system available to interact with consumers must tell each consumer, before or at the time of interaction, that they are interacting with an AI system. The disclosure must be clear, conspicuous, and written in plain language, and it may not use a dark pattern (Bus. & Com. Code § 552.051). That duty applies even when a reasonable consumer would already know they are talking to an AI system (Bus. & Com. Code § 552.051(c)).
If an AI system is used in relation to health care service or treatment, the provider must give that disclosure to the recipient (or their personal representative) by the date the service or treatment is first provided, with a limited emergency exception (Bus. & Com. Code § 552.051(f)).
Private companies that are not governmental agencies and are not providing health care service or treatment do not sit under the general consumer-interaction disclosure rule in § 552.051. They still sit under the intentional-use bans and the Attorney General’s investigative powers below.
Several intentional-use bans.
A person may not develop or deploy an AI system in a manner that intentionally aims to incite or encourage a person to commit physical self-harm (including suicide), harm another person, or engage in criminal activity (Bus. & Com. Code § 552.052).
A person may not develop or deploy an AI system with the sole intent for the system to infringe, restrict, or otherwise impair rights guaranteed under the United States Constitution (Bus. & Com. Code § 552.055).
A person may not develop or deploy an AI system with the intent to unlawfully discriminate against a protected class in violation of state or federal law. A disparate impact alone is not enough to show that intent (Bus. & Com. Code § 552.056). Certain insurance entities and federally insured financial institutions have compliance paths spelled out in that section.
Government-specific limits.
Governmental entities face extra limits on social scoring systems and on using AI to uniquely identify a person from biometric data or scraped public images without consent when that gathering would infringe constitutional or legal rights (Bus. & Com. Code §§ 552.053, 552.054).
Sexually explicit and child-related content.
The Act prohibits developing or distributing AI systems with the sole intent of producing or distributing certain illegal visual material or deepfakes, and prohibits intentionally developing or distributing systems that engage in certain sexual text conversations while impersonating a child (Bus. & Com. Code § 552.057).
The Texas AG publishes the same list on its Consumer AI Rights page.
Who enforces it, and what a complaint can trigger
The attorney general has exclusive authority to enforce Chapter 552, except where another state agency may add sanctions after an AG finding and recommendation (Bus. & Com. Code §§ 552.101, 552.106). There is no private right of action under the chapter (Bus. & Com. Code § 552.101(b)).
Consumers can submit complaints through an online mechanism on the AG’s site (Bus. & Com. Code § 552.102; Consumer AI Rights). The Act directed the AG to post that mechanism by September 1, 2026 (HB 149, Section 8). After a complaint, the AG may issue a civil investigative demand. The demand can ask for a high-level description of the system’s purpose, intended use, deployment context, and benefits; the types of data used to program or train it; input and output categories; performance metrics; known limitations; and post-deployment monitoring and user safeguards (Bus. & Com. Code § 552.103).
If the AG finds a violation, it must give written notice identifying the specific provisions. It may not sue before the 60th day after that notice if the person cures and submits a written cure statement with supporting documentation and the policy changes needed to reasonably prevent further violation (Bus. & Com. Code § 552.104).
Civil penalties for uncured violations run from $10,000 to $12,000 per curable violation (or breach of a cure statement), $80,000 to $200,000 per uncurable violation, and $2,000 to $40,000 per day for a continued violation (Bus. & Com. Code § 552.105; Texas AG overview). The AG may also seek injunctive relief, attorney’s fees, and investigative expenses. The AG may not bring an action to collect a civil penalty for an AI system that has not been deployed (Bus. & Com. Code § 552.105(f)).
The statute also sets out defenses, including discovery of a violation through testing or red-teaming, and substantial compliance with the latest NIST “Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile” or another nationally or internationally recognized AI risk management framework paired with an internal review process (Bus. & Com. Code § 552.105(e)). Those are statute facts. They are not a Hoyack certification claim.
Sandbox and council (context, not the day-to-day checklist)
TRAIGA also creates an AI regulatory sandbox at the Department of Information Resources and a Texas Artificial Intelligence Council (Bus. & Com. Code chs. 553, 554; enrolled bill summary). The sandbox can give a participant limited testing protection for up to 36 months, with possible extension for good cause (Bus. & Com. Code § 553.053). The Chapter 552 Subchapter B requirements may not be waived inside the sandbox (Bus. & Com. Code § 553.051(e)). Those programs matter for state policy and for teams that want limited testing protection. They are separate from the day-to-day disclosure and prohibition rules above.
What this means for a Texas ops or engineering lead
You do not need a new slogan. You need a clear inventory.
- List every AI system your company develops or deploys that could fall under Bus. & Com. Code § 551.001.
- Note which ones interact with Texas consumers, and which touch health care service or treatment.
- Record purpose, intended use, deployment context, training-data types, inputs, outputs, known limits, and post-deployment monitoring. Those are the categories the AG can already ask for under Bus. & Com. Code § 552.103.
- Put disclosure language where Bus. & Com. Code § 552.051 applies, in plain words, without dark patterns.
- Check intentional-use bans in Bus. & Com. Code §§ 552.052 to 552.057 against how your tools are built and sold.
- Escalate open questions on scope or intent to whoever owns compliance decisions for your company.
TRAIGA’s private-sector bans mostly turn on intent. That does not make documentation optional. A complaint can still force you to produce the high-level record in Bus. & Com. Code § 552.103 on a short clock.
Sources (primary)
| Source | Link | Checked |
|---|---|---|
| HB 149 enrolled text (HTML) | https://capitol.texas.gov/tlodocs/89R/billtext/html/HB00149F.htm | HTTP 200, Nov 3, 2026 |
| HB 149 enrolled PDF | https://capitol.texas.gov/tlodocs/89R/billtext/pdf/HB00149F.pdf | HTTP 200, Nov 3, 2026 |
| Enrolled bill summary | https://capitol.texas.gov/billlookup/BillSummary.aspx?Bill=HB149&LegSess=89R | HTTP 200, Nov 3, 2026 |
| Bill history | https://capitol.texas.gov/BillLookup/History.aspx?LegSess=89R&Bill=HB149 | HTTP 200, Nov 3, 2026 |
| LRL bill details | https://lrl.texas.gov/legis/billsearch/BillDetails.cfm?billNumberDetail=149&billtypeDetail=HB&legSession=89-0 | HTTP 200, Nov 3, 2026 |
| Texas AG Consumer AI Rights (TRAIGA overview + complaint route) | https://www.texasattorneygeneral.gov/consumer-protection/file-consumer-complaint/consumer-ai-rights | HTTP 200, Nov 3, 2026 |
