UVU students, faculty demo GovSense tool to automate Utah agencies' privacy obligations

Utah Valley University computer science faculty and students on Friday demonstrated a proof-of-concept tool they say will help state and local agencies meet data-privacy duties under Utah law.

The demo, presented by Dr. George Rudolph, chair of UVU's computer science department, and Caitlin Stratton, a UVU undergraduate, showed GovSense (DataGov AI), an AI-assisted knowledge base and workflow system hosted by the Herbert Institute and supported by UVU's College of Engineering and Technology. The presenters said the tool is intended to guide agency staff through routine privacy tasks including designating a chief administrative officer (CAO), adopting retention schedules and generating privacy notices.

The system's dashboard, Stratton said, maps to "one of the 21 required privacy practices as defined by the Office of Data Privacy" and provides step-by-step tasks for each practice. "All I have to do is click on an employee from the drop down, and then once I've chosen them, designate them as the CAO. The designation is then logged and sent to DARS to be notified, and the system marks that this legal requirement as complete," Stratton said during the demo.

Stratton showed a privacy-notice generator that, she said, accepts plain-text prompts about an agency's activity (for example, a grant program) and returns a draft notice that fills in required elements such as legal authority, retention schedule, data classification, purpose of collection, sharing details and contact information. "The app then produces a complete and fully compliant privacy notice," she said. The presenters noted the demo uses a combination of a knowledge base of codes and statutes and an LLM-style generation layer; the slide deck cited Amazon Web Services for cloud hosting but said other providers and enterprise databases (for example Oracle) could be used in deployment.

Rudolph framed GovSense as a means to standardize protocols and reduce administrative errors while preserving agency responsibility. "One of the big picture ideas is to transform the way that we do data governance by standardizing protocols and policies so that everyone from the smallest town or agency who's under resourced to large agencies across the state can perform their duties and comply with legal requirements and ethical requirements," he said. He also emphasized that policy and legislation should be technically feasible: "Any policy that we make going forward or any legislation is actually technically possible so that we're not legislating things that people can't do."

Presenters repeatedly described the project as a work in progress and a proof of concept built by students under faculty direction. Stratton said generated drafts can be edited, saved in an inventory, or exported as PDFs for distribution. "This is a working proof of concept of how responsible human centered AI can reduce administrative burden, ensure compliance, and build public trust in government operations," she said.

Presenters said the tool is intended to produce audit-ready outputs — for example, automated annual reports showing an agency's compliance posture — and to support decentralized data sources while offering centralized access to policy guidance and required elements. They did not announce procurement, deployment timelines, or formal agreements with state agencies during the session.

The demonstration drew applause from attendees and was followed later in the program by a panel that discussed statewide priorities around standardization, retention schedules and balancing privacy with transparency.