Harris County commissioners were set to discuss an updated artificial intelligence policy on Wednesday, but the item never made it to a vote. The agenda item was pulled, and it's unclear when commissioners will revisit the proposal — leaving open questions about how the county plans to protect residents' personal information as it leans further into AI-powered services.
The proposal, which has not been approved, would replace the county's current 2024 AI policy and strengthen countywide cybersecurity, data privacy, responsible-use and oversight standards, according to Click2Houston. Harris County first introduced generative AI guidelines in October 2023 before approving a countywide responsible-use policy on Nov. 12, 2024. That existing policy covers employees, contractors and anyone else using AI on the county's behalf, addressing responsible use, security, privacy, oversight, and the purchase and operation of AI systems.
A Broader Policy Years in the Making
In January, commissioners directed the county's Universal Services Department to draft a more comprehensive policy, and officials have recommended May 2027 as a target for rolling it out, giving time for employee training and guidance, per the same Click2Houston report. In March, commissioners approved $425,000 in supplemental funding for contractor resources supporting AI compliance, governance and implementation — though that figure does not represent the county's total AI spending.
Until a new policy is adopted, Harris County continues to operate under its existing 2024 framework while officials decide when to bring the update back for a vote. The county is already weighing an expanded footprint for AI in government, with potential benefits like faster responses, shorter wait times and more assistance in different languages, the outlet reports. Separately, GovTech notes the county is rolling out projects touching language access, customer service and emergency preparedness, building on an innovation lab it launched in 2024 to explore practical AI uses.
Experts Flag Privacy, Bias and Accountability Risks
Dr. Dorothy Jiang, an associate professor of information systems and co-director of the Human-Centered AI Institute at the University of Houston's Bauer College of Business, told Click2Houston that privacy, bias and inaccurate information rank among the biggest risks of expanding AI in government. AI systems, she said, can reproduce biases baked into their training data and generate convincing but incorrect answers. Jiang recommended that county data only go into county-approved tools, with each tool seeing only the information it needs — nothing more.
She also recommended that contracts bar vendors from using residents' information to train their own AI models and require that data be deleted once a contract ends. Jiang said residents should be told when AI is involved in decisions that affect them and should have a straightforward way to challenge incorrect information. If a county department uses an AI tool, she said, that department is still responsible for the outcome — the same as if an employee had made the decision.
RaSheem Barnett, CEO of consulting firm Emagine Reality, which focuses on operations, AI and automation, echoed several of those concerns. He recommended limiting access to sensitive information to authorized employees and spelling out in contracts exactly where information goes, who can see it, and whether it can be used to train AI systems. Barnett also recommended collecting only the information necessary for a given task, excluding unnecessary sensitive data like Social Security numbers, and giving a person the authority to review and override AI recommendations in consequential decisions. Residents, he said, should have a clear path to county employees who can resolve problems.
Weighing Costs Before Committing
Barnett warned against buying expensive AI systems before clearly identifying the problem they're meant to solve, recommending that officials evaluate a system's full costs and measurable benefits before deciding whether it actually saves taxpayer money. Both experts framed AI as a tool meant to assist government employees rather than replace human accountability — though the article notes these recommendations are suggestions, not necessarily requirements baked into the county's proposed policy.
Jiang added that AI can assist with routine government paperwork, freeing county employees to focus on cases that genuinely need a human touch, and could help speed services, reduce paperwork and give staff more time for complex cases. She also recommended independent audits of whatever safeguards the county ultimately adopts.
Other County AI Projects Already Underway
The policy debate comes as Harris County pursues other AI applications outside the still-pending governance update. County leaders approved a two-year pilot using AI to speed up permitting for new construction, according to Houston Public Media, after staff research found that few counties nationwide use AI for permitting. The tool is meant to serve as an automated completeness screening before applications go up for review in front of an actual person, with county staff still responsible for conducting the final review and approving new permits. The program is supposed to cut permit processing time by 50%.
The push for oversight isn't unique to Harris County. A UserEvidence survey cited by Smart Cities Dive found that only 19% of state and local government IT professionals reported an active AI oversight framework, and just 20% said such a framework existed at all, even as 77% said their employer was using or piloting generative AI tools like Microsoft Copilot, ChatGPT or Google Gemini. The same survey found that 35% of respondents said their organizations were loosening security or compliance restrictions to enable AI access, while nearly 45% reported an increase in AI-related security incidents over the prior year.
How Residents Can Weigh In
For now, residents who want a say in how the county handles their data have a few avenues available. Harris County Commissioners Court posts upcoming meetings, agendas and public comment opportunities on its website, giving residents a chance to ask which departments use AI, how personal information is protected, and who is responsible when something goes wrong. Advocates can also push for clearer employee contact options, error-reporting channels and human review of AI-driven decisions as the county works out when — and whether — it will bring the updated policy back for a vote.
