Get AI Briefings, Transcripts & Alerts on Local & National Government Meetings — Forever.
Eureka council adopts employee AI policy with new limits on image uploads and role-based use
Loading...
Summary
The Eureka City Council adopted an updated employee artificial intelligence policy that requires training, narrows permitted uses to staffers' job responsibilities, and bars uploading photos or artwork into AI systems without written consent; the vote was 5-0.
Eureka — The City Council on April 21 adopted a new employee artificial intelligence policy (Policy 1.14) that tightens transparency and limits how staff may use generative tools. The policy requires departmental authorization and training, disallows agentic AI without prior approval, and adds a prohibition on inputting photos, videos or artwork into AI systems without express written permission.
Council members and staff said the revisions reflect concerns raised during an ad hoc committee review. Finance Director Millar said the draft drew heavily from a Sonoma County example but was tailored by Eureka staff to add local safeguards. "Language was added in the limited use section to disallow agentic AI use without prior approval," Millar told the council.
Councilmember Katie Moulton, who led the ad hoc discussions, successfully pushed for two specific additions: a rule tying AI use to an employee's job description so staff cannot rely on AI for tasks they cannot personally verify, and a written-consent rule before staff upload photographs or artwork depicting people or created by outside artists. "Don't use AI for tasks outside or above one's professional expertise as depicted in their job description," Deputy City Attorney Holtz proposed for the "do's and don'ts" section; council adopted similar language.
Public commenter and technologist Nick Sarris urged stronger disclosure rules and environmental caution in selecting tools. "It might be smart to require something like a prompt disclosure," Sarris said, recommending more transparency about when AI is used.
Council members also discussed criteria for when a use is significant enough to require a disclosure, noting that if AI becomes ubiquitous the label could lose meaning. The final policy uses a "significance" threshold to avoid stamping every routine spell-check with an AI notice while requiring disclosure for substantive contributions.
The council voted to adopt the policy as amended, 5-0. Councilmembers instructed staff to return with training materials and to revisit the policy as tools and guidance evolve.
