Introduction
The Texas Responsible Artificial Intelligence Governance Act (TRAIGA) [1] [2], signed into law by Governor Greg Abbott on June 22, 2025 [2], represents a significant step forward in AI regulation within Texas, a state known for its conservative stance. This legislation introduces a comprehensive framework for managing high-risk AI systems, focusing on outcome-based restrictions and stringent disclosure requirements. TRAIGA’s enactment positions Texas as a key player in the national AI governance landscape [3], contrasting with more prescriptive approaches in other states [3].
Description
On June 22, 2025 [2], Texas Governor Greg Abbott signed the Texas Responsible Artificial Intelligence Governance Act (TRAIGA) into law [4], following significant amendments in the Texas legislature [2]. This legislation marks a pivotal advancement in comprehensive AI regulation within a traditionally conservative state, introducing a targeted regulatory framework for high-risk AI systems [3]. TRAIGA shifts from extensive compliance obligations to outcome-focused restrictions [3], establishing stringent disclosure requirements for AI developers and deployers [1], while outlining specific prohibited uses of AI. Key prohibitions include the development and deployment of AI systems for behavioral manipulation [2], unlawful discrimination [2] [3], constitutional infringement [2] [3], and the creation or distribution of harmful content, such as child pornography or deepfakes [2]. The Act emphasizes intent as a critical element for liability [3], protecting developers from misuse by third parties while holding accountable those with malicious intent [3]. Notably, it clarifies that “disparate impact” alone does not demonstrate discriminatory intent [3], aligning with federal policy and aiding businesses in compliance [3].
TRAIGA establishes civil penalties for violations [1], with fines ranging from $10,000 to $200,000 depending on whether the violation is curable or ongoing [2]. Effective January 1, 2026 [1], the Act applies to individuals and entities conducting business in Texas [1], including government entities [2], while explicitly excluding hospital districts and higher education institutions [1]. The Texas attorney general has exclusive enforcement authority and will establish a consumer complaint reporting mechanism, allowing for civil investigative demands for suspected violations [2]. Consumers must be informed when interacting with AI [1], necessitating clear and conspicuous disclosures [1].
Additionally, TRAIGA introduces a regulatory sandbox program managed by the Department of Information Resources [2], allowing developers to test innovative AI applications under relaxed regulatory conditions for a period of 36 months, provided they submit quarterly performance reports and risk mitigation measures [3]. The Texas Artificial Intelligence Advisory Council [1] [2], composed of seven members appointed by state leaders [2], will conduct AI training programs and advise the legislature on AI policy [2].
In Illinois [4], a bill was passed that prohibits AI chatbots from functioning as mental health therapists [4], ensuring that only licensed professionals can provide therapy services [4]. In California [4], several AI-related bills are progressing through the legislative process [4]. The AI Abuse Protection Act aims to subject AI-generated content to existing right of publicity laws and criminal impersonation statutes [4]. The AI Transparency Act mandates large online platforms to label AI-generated content and allows users to apply digital signatures to authentic material [4]. Another bill [4], CalCompute [4], seeks to protect whistleblowers in the AI sector who report risks associated with foundational AI models [4].
Furthermore, the Companion Chatbots bill requires AI platforms to remind minors that chatbots are not human [4], while the Human Oversight of AI in Critical Infrastructure bill mandates human oversight for AI systems controlling critical infrastructure [4]. The AI Copyright Protection Act is also under consideration [4], focusing on copyright issues related to AI-generated content [4]. Additionally, a proposal for a 10-year moratorium on state-level AI regulations was passed by the Texas House of Representatives on May 22, 2025 [1], which [1] [2], if enacted [1], would preempt TRAIGA and other active state AI-related legislation [1]. The passage of TRAIGA positions Texas as a significant player in the national AI governance landscape [3], contrasting with more prescriptive regulatory approaches in other states [3], and contributing to a fragmented regulatory environment that may prompt calls for federal preemption or a unified national approach. Ultimately, the effectiveness of the Act will depend on the ability of traditional regulatory frameworks to adapt to rapidly evolving AI technologies that impact human decision-making and agency [3]. While TRAIGA addresses intentional harmful uses [3], it raises complex questions regarding AI systems that influence decision-making without explicit intent [3], highlighting the challenges of applying existing legal frameworks to emerging technologies [3].
Conclusion
The enactment of TRAIGA underscores Texas’s proactive stance in AI governance, setting a precedent for other states. By focusing on outcome-based restrictions and stringent disclosure requirements, the Act aims to balance innovation with accountability. However, the fragmented regulatory landscape across states may lead to calls for a unified national approach. The effectiveness of TRAIGA will ultimately depend on its adaptability to the fast-evolving AI landscape and its ability to address both intentional and unintentional impacts of AI systems on human decision-making and agency.
References
[1] https://natlawreview.com/article/traiga-key-provisions-texas-new-artificial-intelligence-governance-act
[2] https://www.lw.com/en/insights/texas-signs-responsible-ai-governance-act-into-law
[3] https://natlawreview.com/article/texas-enacts-responsible-ai-governance-act
[4] https://www.transparencycoalition.ai/news/ai-legislative-update-june-27-2025
