The Texas Office of the Attorney General switched on its consumer AI-complaint portal on September 1, 2026, meeting the statutory deadline written into the Texas Responsible AI Governance Act. TRAIGA took effect on January 1, 2026, but until this week it was a law without a doorbell. Now any Texas resident can file online.

The reach is what matters. TRAIGA defines an “AI system” as any machine-based system that infers outputs from inputs, which is a definition broad enough to sweep in the chatbot on a plumber’s website, the personalization engine behind a boutique’s email list, and the AI-drafted cold outreach a five-person B2B shop sends every Tuesday morning. There’s no revenue threshold and no size floor.

Enforcement sits exclusively with the AG, per Latham & Watkins; there’s no private right of action. But the portal itself is the mechanism. Duane Morris flagged the September deadline as the moment when enforcement activity was expected to accelerate, and the OAG page now lists prohibited uses including intentional manipulation and unlawful discrimination. Disparate impact alone doesn’t prove intent, Duane Morris and Haynes Boone note, but a complaint can trigger a civil investigative demand seeking training data, inputs, outputs, and performance metrics.

The penalty math concentrates the mind. Curable violations run $10,000 to $12,000 each, with a 60-day cure window after AG notice. Uncured violations run $80,000 to $200,000. Continuing violations can reach $40,000 per day.

The cheapest defense is documentation: written AI-use policies, NIST AI RMF alignment, and a paper trail showing intent. Small businesses that have been treating AI outreach as a productivity hack now hold a compliance posture whether they’ve written one down or not.

Texas isn’t moving alone. The FTC’s AI accuracy-suppression rule is closer to law, and the agency has already put operators on notice that undisclosed AI pricing could be deception. The complaint form is simply the first one that talks back.

Sources

Sources