Texas enacts Responsible AI Governance Act (TRAIGA)
The law bars specific harmful uses of AI, such as manipulating behaviour or discriminating unlawfully, rather than regulating systems by risk category as the EU and Colorado do.
- Government & policy
- Notable
Texas Governor Greg Abbott signed the Texas Responsible Artificial Intelligence Governance Act, making Texas the second US state, after Colorado, to enact a broad AI-specific statute rather than regulating AI only through existing sector law. The bill had started life in 2024 as a comprehensive, EU AI Act-style framework classifying systems by risk category; by the time it reached the governor’s desk, legislators had narrowed it substantially.
As enacted, TRAIGA does not regulate AI systems by risk tier. Instead it prohibits specific uses: developing or deploying AI to manipulate human behaviour in ways that cause harm, to unlawfully discriminate against a protected class, to produce child sexual abuse material or unlawful deepfakes, or to infringe constitutionally protected rights. It creates a regulatory sandbox in which companies can test AI products with reduced compliance obligations under state oversight, and establishes a Texas Artificial Intelligence Advisory Council to monitor the technology’s development. The law takes effect on 1 January 2026.
Legal commentary framed the shift from the 2024 draft to the enacted version as evidence of an industry lobbying push that succeeded in most US states during the period: intent-based prohibitions on specific harmful outcomes, rather than upfront risk assessment and audit obligations tied to a system’s technical characteristics. That distinction — outcome-based versus risk-tiered regulation — became one of the clearest fault lines in US state AI law through 2025, with Colorado’s earlier, more prescriptive act facing its own delays and amendments in the same period.