Cyber Command Envisions AI Agents Across State Government

Texas Cyber Command (TXCC) is exploring the use of artificial intelligence agents to continuously monitor government networks, identify vulnerabilities and strengthen cybersecurity defenses, with a long-term vision of assigning an agent to every state agency and public critical infrastructure owner and operator.

Cyber Chief TJ White outlined the concept Tuesday during the virtual Rubrik Public Sector Summit, describing how agentic AI could help the newly established agency expand its cybersecurity capabilities and respond to increasingly automated threats.

The proposed agents would operate continuously as part of human-led cybersecurity teams, assessing assigned networks for vulnerabilities, testing software in designated environments and recommending cost-effective security measures.

“With more than 150 state agencies in the Texas government to protect, we envision creating agentic work roles as part of our human-led and operated teams,” White said.

The command would first need to develop, build and test the model before considering broader deployment.

“So after we develop, build and test this model, assuming success, we envision that we can assign an agent to every state agency and every public critical infrastructure owner and operator,” White said.

White did not provide an implementation timeline, cost estimate or procurement details for the proposed system.

The concept is part of the command's broader examination of how AI can support cybersecurity operations. White divided potential applications into conventional and unconventional approaches.

Conventional uses include automated penetration testing, vulnerability assessments, patch management and software code analysis. More experimental applications could include AI-generated digital replicas of government networks, simulated environments for testing cyber attacks and autonomous agents monitoring traffic entering and leaving protected systems.

White said AI could help the command extend its reach despite workforce limitations, particularly as cyber adversaries increasingly use the technology to identify vulnerabilities and conduct attacks at greater speeds.

“The only way to pace and defeat an AI-enabled adversary is with AI capability of our own,” White said. “You cannot defend against an autonomous machine-speed attack with human-speed defense.”

The command's exploration of AI capabilities comes as the agency seeks about $235 million in exceptional-item funding through its Legislative Appropriations Request for the 2028-29 biennium.

Of that amount, $42 million is requested for AI security and quantum readiness, including $25 million for approved AI model access, computing resources, automation and engineering. Another $10 million would support statewide AI system security and assurance, while $6.5 million would support cryptographic discovery and post-quantum readiness.

The funding request describes a broader effort to establish statewide capabilities for discovering, assessing, testing and securing emerging technologies. It does not identify White's proposed AI agent deployment model as a separately funded initiative.

As the agency considers greater use of autonomous technologies, White emphasized that safeguards would need to be developed and tested alongside the AI agents.

The proposed systems would be expected to operate within the command's operational frameworks, assigned missions and regulatory obligations, with human personnel maintaining oversight.

The challenges of managing AI agents have also surfaced among other Texas technology leaders.

During September's State of Technology Texas Industry Forum, Department of Public Safety CIO Jessica Ballew identified agent identities, system access and release management as emerging governance concerns. Ballew said agencies were still determining how best practices should be established and applied as environments with multiple AI agents become more common.

White also pointed to the Texas Department of Information Resources' role in assessing safe AI adoption across state government. He highlighted three initiatives being pursued by the department: an AI code of ethics for state and local governments, a public-sector AI sandbox program and mandatory annual AI awareness training for state and local government employees.

He distinguished those broader government initiatives from Texas Cyber Command's operational focus on preventing and responding to cybersecurity incidents.

White said the command must also account for potential risks associated with autonomous AI systems, including models taking actions outside their intended environments. He called for working with AI developers to establish safeguards while preparing defensive capabilities in case those protections fail.

“We see agentic AI as a way of giving our finite number of personnel additional teammates in the fight, and the ability to scale cyber defenses across the state,” White said.