Texas agencies outline AI rules, sandboxes and required training for public institutions

State officials told the LTAC that the Texas Legislature has enacted several bills this year that create a compliance framework for AI in public institutions and give DIR a central administrative role.

Jenny Holscher of the Department of Information Resources (DIR) described four bills she said will shape how colleges, universities and state agencies adopt AI. "House Bill 149" (the private‑sector focused responsible‑AI bill) spells out prohibited uses and requires public‑facing disclosure when the public interacts with AI. Holscher said DIR is preparing inventory questionnaires and guidance to help agencies comply with the statute's disclosure and inventory requirements.

Holscher said Senate Bill 1964 is the primary public‑sector AI statute. It directs DIR to publish an AI code of ethics and minimum standards for risk management of so‑called "heightened scrutiny" AI. That rulemaking is under way and a draft code and standards were published for public comment in the Texas Register on Nov. 7, she said. "The minimum standards that are in that rule are limited to the heightened scrutiny AI," Holscher said, emphasizing that many AI deployments will not meet the threshold for heightened scrutiny.

SB 1964 also established a Texas AI Systems Advisory Board and a public‑sector AI sandbox that DIR will operate; Holscher said the governor recently announced appointees to that advisory board and that DIR is standing up an AI division and job postings for the new unit.

House Bill 2818 directs DIR to create a division to modernize legacy state systems using AI (with associated funding), and House Bill 3512 requires annual AI training for state employees; DIR will offer a free training option and a process for certifying other trainings.

A recurring theme in committee discussion was vendor transparency and procurement. Members warned that many third‑party learning platforms add AI features after deployment and that institutions do not always receive formal notice. Holscher said contract terms will be revised to require vendors to comply with NIST AI risk‑management frameworks and to disclose AI features to state customers.

The DIR representative encouraged higher‑education leaders to review the draft code and to participate in public comment. She also said DIR plans an AI literacy course and expects to publish certified training materials early next year.

Committee response: members asked whether higher education will be represented on the Texas AI advisory boards and requested clarity on the disclosure format for AI features embedded in learning tools. Holscher said public‑sector board meetings will be open to the public and that she will confirm whether higher education has a seat on the public‑sector advisory board and circulate the membership list.