SEPTEMBER 24: In this photo illustration, ChatGPT opened on a computer screen is displayed on September 24, 2026 in Melbourne, Australia. The AI agent “hacked a statistics portal containing private data from Australia’s universal healthcare scheme, Medicare,” local media reports said, which experts believe is “the world’s first known AI breach of a government system.” (Photo by Asanka Ratnayake/
AUSTIN, Texas —Some Texas lawmakers are moving to regulate artificial intelligence in the upcoming legislative session with bills that could put new safeguards in place.
One bill, HB 1201, filed by State Rep. Salman Bhojani (D-Euless), would require developers or frontier AI models to report “rogue” artificial intelligence incidents to the Texas Department of Information Resources.
ALSO | Texas Comptroller proposes new tax rules for cloud computing, website hosting and AI
The bill defines the incidents as events in which an AI model accesses a computer, computer system or computer network without the owner’s effective consent or acts to override a human oversight mechanism.
It also covers incidents in which a model, without the owner’s effective consent, modifies, deletes or discloses data or takes an action affecting an information system.
The measure says it would also apply when a model takes action beyond its instructions and causes serious bodily injury, substantial property damage, substantial loss or disruption of critical infrastructure, substantial disruption of a government service or a substantial compromise of confidential information or an information system.
The bill says a frontier developer would have to report an incident within 24 hours if it poses an imminent threat of substantial harm to the health or safety of one or more people. Other incidents would have to be reported within 72 hours.
Bhojani has filed 13 bills related to AI oversight ahead of the legislative session that begins in January. The bills would regulate how AI is used by state agencies, schools and health care providers.
In a statement to CBS Austin on Wednesday, Bhojani wrote:
Artificial intelligence changes every day, but the Texas Legislature only meets every two years. We can’t wait until something goes wrong to put safeguards in place.
AI is already shaping how Texans get health care, apply for jobs and deal with their government, and these systems are only getting more powerful. When an AI system starts acting outside the limits it was designed for, the public deserves to know before anyone gets hurt. That’s why I filed legislation requiring developers to report incidents involving rogue AI.
Cybersecurity and artificial intelligence experts argue that the technology may be difficult to regulate, but agree a mandatory reporting system is a step in the right direction.
“I think it’s a step in a direction where we’re trying to feel our way through what kind of regulation needs to be put in place,” said Bill Hefley, clinical professor emeritus at the University of Texas at Dallas.
ALSO | TCEQ rejects bid to raise inspection prices in Austin, El Paso
Hefley said he had questions about why only developers would be required to make reports of rouge systems, and if the Department of Information Resources was the best agency to process reports.
“There’s some questions about, will the bill be modified before final passage? And I expect it might, but it does give a way to start collecting some of that accident or incidents data,” he said.
Guofei Gu, a professor of engineering at Texas A&M University, said proactive regulation from the legislature would be much more difficult than the requirement of reporting them after the fact.
“The legislation, I think, is a great step. However, it’s not going to prevent attacks from happening,” Gu said.
Supreeth Shastri, a computer science professor at the University of North Texas, said the uncertain future of artificial intelligence adds a layer of difficulty when passing meaningful legislation.
“We are all in the, sort of, same boat. Maybe some boats are a little bit ahead of others, but we truly do not grasp this very well. Any time you want to draft regulation, even the one that I kind of like, I generally tend to be supportive of accountability once something has happened. Even those ones could be full of problems, right?” Shastri said.
If passed by the legislature, the bills would go into effect September 1, 2027 and the reporting system for rogue AI incidents would have to be in place by December 1, 2027.
3
NEWStudent injured with scissors during altercation at Akins Early College High School3
San Marcos police step in to help Kyle officers attend fallen colleague’s funeral17
Terry Black’s Barbecue opens first San Antonio restaurant3
Eanes ISD trustees reject 4 proposals to limit out-of-district transfers
