From the Local Government edition of October 31, 2026
Island County, Washington's generative AI policy for a 27 Oct roundtable would require staff to disclose prompts, responses and tool for any AI-assisted record
Island County, Washington · County administration and records · Island County, Washington, United States
- Publisher
- Island County Agenda Center
- Original publication
- 27 October 2026 (roundtable agenda packet)
- Source retrieved
- 2026-11-01
- Event date
- 2026-10-27
What happened
Island County's generative AI use policy, on the agenda for a 27 October roundtable of elected officials and department heads, goes further than most county policies on records. For any document, record or data produced with GAI help, staff must disclose that GAI was used, the prompts, the responses and the tool. Staff must also record that they verified the output. The policy limits use to IT-approved tools, bars PII, HIPAA, 42 CFR Part 2 and CJIS data without director consultation, forbids GAI from directly modifying or deleting county data, and notes that prompts and outputs may be public records under Washington law.
Why it matters
Why now: the 27 October agenda shows a mid-sized county putting a prompt-level disclosure rule in front of its elected officials. It is a concrete example of public-records law shaping AI controls in a Washington county.
Evidence and measured results
Policy text attached to an official county agenda. No adoption record or implementation data.
Limitations and uncertainty
Adoption status unknown; agenda-item evidence only.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-31; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- A county must keep a public-records trail of AI prompts and outputs without overloading staff.
- Stakeholders
- IT Director
- Elected officials
- Department directors
- Public records officer
- Discovery questions
- Has the policy been adopted, and with what changes?
- Which GAI tools are approved?
- Where will prompts and responses be stored?
- How will records requests for prompts be handled?
- Value hypothesis
- Bounded: standard logging and templates could make prompt disclosure routine rather than manual.
- Potential engagement
- Records-and-logging design for an adopted county GAI policy.
- Unsupported claims
- The policy is adopted
- Staff currently fail to disclose
- Any tool meets the rule today
Pre-sales engineering
Role takeaway
- Fit
- Counties in states where prompts and outputs are public records.
- Architecture and integration
- Enterprise GAI tools with audit logging linked to document or record management.
- Prerequisites
- Approved tool list
- Records retention schedule for AI interactions
- Logging capability in chosen tools
- Constraints
- Washington Public Records Act
- No direct GAI modification of county data
- Sensitive data categories barred
- Security
- Block PII, HIPAA, 42 CFR Part 2 and CJIS data in GAI tools unless a director approves.
- Proof of value
- Run one department's AI-assisted documents through the disclosure process and time it
- Test retrieval of prompts for a mock records request
Delivery
Role takeaway
- Work
- Finalize the policy, publish approved tools, set up disclosure and retention, and train users.
- Dependencies
- Policy adoption
- Tool logging features
- Records officer guidance
- Ownership
- County IT Director, with department directors for sensitive data decisions.
- Skills and adoption
- Staff learn to verify output and record prompts and tool use.
- Governance checkpoints
- Adoption by the county
- Quarterly disclosure spot checks
- Annual tool list review
- Acceptance criteria
- Proposed: disclosure recorded for sampled AI-assisted documents
- Proposed: prompts retrievable within the records-request deadline
- Risks
- Disclosure becomes a formality
- Retention costs grow
- Embedded AI summaries go unrecorded
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Keeping prompts, responses and tool names for each AI-assisted record needs logging that can be tied to the record, which chat tools do not do by default.
Governance
Who approves, reviews and stays accountable for outcomes?
Disclosure of prompts and responses plus recorded verification creates an audit trail, but also a records-retention workload for every department.
Security and privacy
What data, permissions and controls need testing?
Named prohibitions on PII, HIPAA, substance-use (42 CFR Part 2) and CJIS data, and a ban on GAI directly changing county data.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
The policy states GAI is not intended to replace any position; verification and disclosure add work for staff.
Procurement
What should contracts, pricing and exit terms secure?
Approved tools should be able to export or retain prompts and responses to meet the disclosure rule.
Operating model
Which teams own the service once it runs?
The IT Director approves tools; department directors decide on sensitive data use.
Publication history
- 2026-10-31Local Government · Issue 564 resources
Stable resource ID: island-county-wa-genai-policy-prompt-response-disclosure-roundtable-20261027