DuPage County releases AI guidelines, plans Copilot pilot and training for staff

Anthony, an Information Technology staff member, presented draft artificial‑intelligence guidelines for DuPage County to the Technology Committee on May 6, 2025, saying the document is intended to give county users guidance on responsible, secure and ethical AI use while allowing the county flexibility to adjust as technology evolves.

Why it matters: the guidelines aim to enable efficiency gains from AI while protecting privacy and county systems. Anthony said the document applies to county end users on county‑owned equipment, networks and technologies and covers generative AI, machine learning, natural language processing, computer vision and robotics.

Anthony described the development process: an AI task force, review of existing public‑sector policies and consultation with Gartner. The guidelines recommend that any new AI initiative or upgrade be reviewed by IT to ensure compatibility, security and ethical compliance. They prohibit "processing of personally identifiable information, HIPAA data, or law‑enforcement sensitive data unless explicitly approved," require that AI‑generated media be clearly labeled, and forbid using personal devices to bypass county safeguards.

Anthony said the county will move toward a closed AI loop for county data: building tools that query only designated county data sources. He also said the IT team plans to pilot a customized Microsoft Copilot and use Copilot Studio to build an in‑house chatbot; the larger rollout is expected after a 6–9 month testing period. He noted licensing costs and said deployment will be targeted to staff with demonstrable need.

On training, Anthony said the county will make role‑specific, annual AI training widely available and strongly encourage elected officials to participate, though he said, "We can't force elected officials to do anything, but we are gonna strongly encourage them to follow that lead." The county has reviewed a free course from Innovate US and built a branded portal to host training resources.

Anthony also presented survey results: staff who said they were "very" or "extremely" familiar with AI rose from 14.6% in 2024 to 26.6% in 2025; active AI usage rose from 15.8% to 26.2%; 74% of staff expressed interest in AI training. Respondents ranked perceived benefits as efficiency (48%), automation of repetitive tasks (40.9%), improved accuracy (39.7%) and analytics (37.6%). Concerns rose slightly (47.4% to 52.3%), led by data privacy and security, job displacement, accuracy and ethical misuse.

Committee members asked about timing, access and whether elected officials would be included in training. Anthony said the guidelines will be issued to staff after the committee provides feedback and that broader piloting of Copilot could be completed by the end of the year, with targeted licenses for early adopters who agree to provide feedback.

Ending: Anthony asked the committee for feedback; staff said they will publish the guidelines, host IT‑led awareness sessions and continue to update the document as the technology and rules evolve.