Field Notes · Ballot-Ready & Civic Office
Using AI in Public Service Without Losing Public Trust: Guardrails for Local Agencies
AI can help agencies answer faster and do more. Without guardrails, it can also make mistakes at scale. These steps protect public trust.
Local agencies are trying AI for everything from answering calls to drafting reports. Done well, it can save time. Done poorly, it can make errors that harm people and erode trust.
Start With the Right Uses
- Good starting points: drafting, summarizing, translating, and searching records.
- Use more care with anything that decides who gets benefits, services, or penalties.
Guardrails That Matter
- A human decides: staff review and own every decision that affects a person.
- Tell people: say when AI is used and how to reach a person.
- Protect data: keep private information out of tools that are not approved for it.
- Test for bias: check results across neighborhoods, languages, and groups.
- Keep records: log what the tool did, so errors can be found and fixed.
Write It Down
A short, public AI policy builds trust. It should list approved tools, banned uses, who is responsible, and how the public can raise concerns.
Dr. Allison Alimayu holds a PhD in biomedical informatics. She is a former elected official and a public interest futurist, and she leads Alimayu Studio.