Texas TRAIGA
HB 149: conditional defence to liability and prohibited-use detection
Updated
Texas HB 149 (Responsible AI Governance Act) takes effect 2026-01-01 and applies an intent-based liability framework to AI development and deployment in Texas. The Act provides a conditional defence to liability under BCC 552.105(e): a defendant may not be found liable if a violation is discovered through the 552.105(e)(2)(D) internal-review discovery route and the defendant substantially complies with the NIST Generative AI Profile.
Asqav can sign submitted agent-action records with ML-DSA. It retains the receipts under the longer of HIPAA six years or any Texas-required period, and emits stable reason codes for the prohibited-use categories of BCC 552.052.
The receipt format is profiled in the IETF Internet-Draft draft-marques-asqav-compliance-receipts.
How Asqav supports HB 149 obligations
| Section | Requirement | Asqav binding |
|---|---|---|
| BCC 552.105(e)(2)(D) | Conditional defence to liability under BCC 552.105(e): violation discovered through an internal review process, with NIST Generative AI Profile substantial compliance. | Audit pack as evidence. NIST AI RMF binding co-applied via the same receipt substrate. |
| BCC 552.052 | Prohibited-use categories: incitement of self-harm, harm to another, criminal activity. | Three stable reason codes on deny receipts: prohibited_self_harm_incitement, prohibited_harm_to_another, prohibited_criminal_activity. |
See the docs for the full mapping and the IETF draft for the binding text.