Pacific Grove directs staff to develop citywide artificial intelligence framework

The Pacific Grove City Council on Sept. 3 unanimously directed the city manager to begin drafting a citywide artificial intelligence (AI) framework and policy for future council review and adoption.

Councilmember (presenter) raised the item to ensure the city “is ahead of the curve” as AI tools proliferate. The staff presentation described three broad AI categories — predictive, generative and perceptive — and outlined risks such as bias, model errors, data privacy and legal liability if systems are used without governance. The staff report recommended a six‑month development period to produce a draft policy and implementation plan.

Council discussion focused on the scope of the policy, anticipated staffing and training needs, and whether work should be led by IT or treated as an enterprise‑level policy affecting multiple departments. Councilmembers emphasized training and practical use cases in addition to restrictions, with one councilmember recommending explicit language making clear whether types of high‑risk uses would be prohibited. Another member urged inclusion of risk mitigation for material damages and financial controls for systems that might touch city finances.

Public commenters expressed a mixture of caution and support. Several speakers urged the council to adopt a policy that both curbs risks and enables staff to use AI tools to increase efficiency. One resident described AI as “frightening,” while another, who works with AI, said proper training and clear guidelines would let staff use the tools effectively while reducing hallucinations and other errors.

The council motion directed the city manager to initiate development of a comprehensive, citywide AI framework and policy. Councilmember Mogensen (city manager) told the council six months was a reasonable timeline to assemble relevant models and resources and to consult League of California Cities resources and other jurisdictions with existing policies.

The direction includes: identifying allowable business uses, training needs, data privacy and security safeguards, guidance on vendor and model selection, identification of prohibited or high‑risk applications, and suggested employee access controls. Staff said they would report back with recommended policies, training plans and potential resource needs.