Texas' Responsible AI Governance Act, known as TRAIGA, has been binding law since January 1, 2026. Seven months in, most businesses using an AI chatbot, an AI-written FAQ page, or an AI-driven support widget still have not looked at whether their site discloses that fact anywhere. That gap matters more this year than it used to, because Texas is not regulating AI in isolation. California's AI Transparency Act took effect the same day. Connecticut signed its own AI and online-safety law in late May, with the first obligations landing October 1. Colorado's AI Act, once expected in February, has been pushed back twice and now will not bind anyone until January 1, 2027.
The result is a patchwork where "is AI regulated in my state" has a different answer depending on the month you ask it. TRAIGA itself is narrower than its name suggests, and it is worth being precise about who it actually reaches before deciding what to change on your site.
Source: OneTrust, "Where AI Regulation Is Heading in 2026: A Global Outlook," captured August 2026.
What TRAIGA actually requires
TRAIGA's blanket "tell the user they are talking to an AI system" mandate applies to state agencies and to healthcare providers using AI in diagnosis or treatment, not to private businesses generally. A private employer using AI in hiring is not required to disclose that to applicants under this law. That narrower scope is easy to miss when a headline says "Texas has an AI transparency law now," and it is worth getting right instead of adding disclosure language your business was never required to carry.
The part of TRAIGA that reaches ordinary businesses runs through a different mechanism: the Texas Deceptive Trade Practices-Consumer Protection Act. TRAIGA folds AI-related deception into the DTPA, so a business that uses AI to mislead consumers about a product's origin, sponsorship, or approval is exposed to the same liability any deceptive practice would create, regardless of industry. Practically, that means a chatbot or AI-generated communication a reasonable person could mistake for a human is the kind of thing regulators and plaintiffs' attorneys will look at first. The Texas Attorney General has exclusive enforcement authority, and TRAIGA sets civil penalties from $10,000 to $200,000 per violation, which can accrue for each day a violation continues.
Where the other three states stand
Figure: Compliance deadline for each state's AI disclosure law, shown in months since January 2025. Texas and California are already past their deadline; Connecticut and Colorado are still ahead.
California's AI Transparency Act, also effective January 1, 2026, sits at a different scale than TRAIGA. It targets "covered providers," meaning generative AI systems with more than one million monthly users that are publicly accessible in California, and it requires them to offer a free AI-content detection tool plus visible and embedded disclosures on AI-generated image, video, and audio content. Most small and mid-sized businesses using a third-party AI tool are not the covered provider under this law, the platform they license from typically is, but that does not mean the law is irrelevant to a business relying on that vendor's promises around content labeling.
Connecticut signed Senate Bill 5, its Act Concerning Online Safety, on May 27, 2026. It is a sweeping bill covering AI companions, employment AI, and provenance requirements, with obligations staggered across three effective dates. The earliest batch, including a requirement that businesses offering AI technology through a subscription give consumers clear written notice of key service terms, lands October 1, 2026. AI companion provisions follow on January 1, 2027, and requirements for covered online platforms follow on January 1, 2028.
Colorado's AI Act has had the roughest path to the calendar. It was originally set to take effect February 1, 2026, then delayed to June 30, 2026, and in May 2026 Governor Polis signed a further delay pushing the compliance deadline to January 1, 2027, while also narrowing the law away from its original risk-assessment framework toward a disclosure-and-transparency model. As of this post's publish date, Colorado's AI Act binds no one yet.
What this means for your disclaimer
The practical takeaway is not that every business needs a Texas-specific AI clause. It is that "AI disclosure" is no longer a single, deferrable line item, it is four different state timelines moving at once, and the two that are live right now, Texas and California, both carry real enforcement teeth attached to deception and misrepresentation rather than to AI use itself. A disclaimer that already tells visitors when content, chat responses, or recommendations on your site are AI-generated or AI-assisted closes the gap that TRAIGA's DTPA provision and California's provenance rules are both, from different angles, aimed at. Our Disclaimer Generator includes AI-use disclosure language you can add to your site now, before Connecticut's October deadline or Colorado's 2027 date turns this into a five-state patchwork instead of a two-state one.
The information in this article is for informational purposes only and should not be construed as legal advice on any matter, and does not create a lawyer-client relationship.
