AUSTIN (KXAN) — Four Texas agencies detailed how they’re currently using artificial intelligence tools in their operations and what they may plan to do with it in the future. They shared this information Thursday during the second meeting of the state’s newly-created AI advisory council, which is tasked in part with taking stock of what state agencies are doing with the rapidly-developing technology.
The seven-member panel heard from people working with the Texas Department of Information Resources, the Texas Department of Transportation, the Texas Workforce Commission and the Teacher Retirement System of Texas. State Rep. Giovanni Capriglione, R-Southlake, serves as co-chair of the advisory council and told KXAN the witnesses were all invited to share testimony Thursday. He shepherded the legislation that created this new panel through the House in the regular legislative session last year.
Anh Selissen, the chief information officer for TxDOT, shared how the agency is moving forward in a “cautiously innovative” way with AI. However, one initiative she discussed is a pilot program for AI incident detection. She said this uses software to create an analysis on areas that are perhaps the biggest safety risks for construction and maintenance. It’s then supposed to lead to faster incident notification and response times as well as improve coverage for any problems that arise.
Selissen said it’s made a difference in receiving quicker notification of incidents on the roads within the Austin district, where this pilot is currently underway. She noted it shortened response times there by five to 10 minutes. The two lawmakers who co-chair the council shared how impressed they were to hear that.
The hope is to expand this AI incident detection software statewide, she added.
Representatives from the Teacher Retirement System told council members they are in the “crawling stage of the use of AI” and have not rolled out anything specifically. However, the agency developed a policy in October related to AI and created a review team who will look into the tools or other opportunities moving forward.
They noted, though, that the agency would never use AI to make decisions, only to instruct or augment information for an actual person to make a decision.
What the law requires
House Bill 2060, which the governor signed into law last year, created the Artificial Intelligence Advisory Council for the state. The legislation’s author, Capriglione, will serve as the council’s co-chair after House Speaker Dade Phelan named him to the role. Meanwhile, Lt. Gov. Dan Patrick named Texas Sen. Tan Parker, R-Flower Mound, as the other co-chair.
Gov. Greg Abbott appointed another four people to serve on the council. That included two Austinites: cybersecurity expert Dean Teffer and attorney John Bash. The others are Mark Stone, the chief information officer at Texas A&M University, and Angela Wilkins, the executive director of the Ken Kennedy Institute at Rice University.
Capriglione told KXAN that the council plans to meet monthly, and at some point they’ll open up the meetings to include testimony from members of the public.
Every Texas agency has until July 1 to turn in what the law states are automated decision systems inventory reports. These will detail “all automated decision systems that are being developed, employed, or procured by the agency,” according to the law.
The AI advisory council will then review these inventory reports from the state agencies and summarize the findings for state lawmakers. The members have until Dec. 1 to meet and craft a report for the state legislature that looks into a number of things, including how state agencies are using automated systems, assessing whether a state code of ethics about AI is needed as well as putting forward policy recommendations. The timing of that is important since it comes after the November general election and ahead of the start of the next session in January.
The law states the advisory council will expire on Jan. 1, 2025, so this will not be a permanent working group unless other legislation is introduced next session to further this focused effort on AI.