Texas Just Wrote the Anti-Illinois AI Hiring Law. Here's Why That Still Matters for Indian HR

Texas’s AI hiring law took effect January 1, 2026, and it reads almost nothing like the state laws we’ve covered before. The Texas Responsible Artificial Intelligence Governance Act (TRAIGA), signed by Governor Abbott in June 2025, makes Texas the third state to regulate AI hiring after Colorado and Utah. But where Illinois, Colorado, and NYC built their laws around candidate disclosure, Texas explicitly didn’t. Its disclosure duty applies to government agencies talking to consumers, not private employers talking to candidates. That gap between “third state to regulate” and “the opposite kind of regulation” is the part worth an Indian HR team’s attention, not the headline.

What TRAIGA actually requires

Two provisions matter for a screening tool. First, TRAIGA makes it unlawful to develop or deploy AI “with the intent to discriminate” against a protected class: race, color, national origin, sex, age, religion, disability. That’s a narrower bar than the disparate-impact standard most other AI-hiring laws use. Unequal outcomes alone don’t establish a violation under TRAIGA the way they can under an EEOC disparate-impact theory or the EU AI Act’s risk-management duties. Second, TRAIGA requires explicit written consent before collecting biometric identifiers, including through AI hiring tools, which does bite voice and video screening vendors specifically, since a voice or video interview can be read as biometric collection depending on how the analysis pipeline works.

Enforcement sits entirely with the Texas Attorney General. There’s no private right of action, meaning a candidate can’t sue an employer directly under TRAIGA the way they increasingly can under some other states’ frameworks. Employers get a 60-day cure period once the AG flags a violation, and the AG’s public complaint portal isn’t live until September 1, 2026, so meaningful enforcement activity likely doesn’t start until well into the second half of this year.

Why “employer-friendly” doesn’t mean “skip it”

It would be easy for an Indian TA team to read all of that (narrow intent standard, no private lawsuits, no disclosure duty, enforcement barely starting) as the one AI-hiring law that’s safe to ignore. That reading misses two things. One, TRAIGA’s reach is broad on paper: it applies to any entity that does business in Texas, sells to Texans, or deploys AI in the state, which pulls in offshore delivery and recruiting operations, not just Texas-incorporated employers. Indian IT services majors and GCCs that screen candidates for Texas-based roles or run recruiting operations touching Texas are inside the law’s scope even if no office sits in Austin or Houston.

Two, and more useful: the biometric-consent requirement is real and specific in a way the discrimination standard isn’t. “Intent to discriminate” is a high bar that’s genuinely hard to trigger with a well-built screening tool. Explicit written consent before collecting a voice sample for analysis is a concrete operational step. Either you have it on file before the interview or you don’t, and it’s the kind of requirement that shows up in a vendor-diligence checklist regardless of how aggressively the AG enforces the rest of the law.

The contrarian take

Every other AI-hiring law we’ve written about, Illinois’s HB3773, Colorado’s rewritten act, NYC’s Local Law 144, the EU AI Act, converges on the same two requirements: tell the candidate AI is involved, and keep a human able to override the outcome. TRAIGA is the first major state law we’ve seen that deliberately skips the first half. That’s not a signal the disclosure trend is reversing. It’s a signal that “which US state’s law applies to us” is the wrong question for an Indian vendor or GCC to be optimizing for, because the answer keeps flipping between two incompatible models depending on which state you’re reading. The stable move isn’t tracking each state’s disclosure stance. It’s building disclosure and human override as defaults regardless of whether the specific state you’re hiring into currently requires them, because Texas is the exception this year, and exceptions in this space have not stayed exceptions for long.

What to actually check before your next US-facing drive

If your screening pipeline touches Texas-based roles or candidates, the concrete item to close isn’t a disclosure banner. TRAIGA doesn’t require one. It’s biometric consent capture: does your AI screening vendor obtain and log explicit consent before a voice or video interview runs, and can you produce that record if the AG’s complaint portal generates an inquiry after September 1? That’s a question worth asking any vendor’s compliance documentation directly, not inferring from a marketing page. Talk to us if you want to see what that consent flow looks like in a live screening pipeline.