Half Moon Bay is writing rules for how its city government uses artificial intelligence. At its Oct. 20 meeting, the City Council heard a report on the city's draft AI policy, the work of an internal AI subcommittee and an AI capability the city already hosts on its own infrastructure. Staff asked for feedback and direction, according to the meeting's agenda item as posted by Coastside Buzz.

What the draft says

The draft centers on safeguards and human review, the San Mateo Daily Journal reported. It was modeled on policies from nearby cities. Its guidelines include IT approval of every new AI tool and a rule that AI never makes major decisions on its own. Staff said they prefer AI programs that run only on city computers, so third parties don't get city or resident data, with several layers of protection on every system.

What AI might do for the city

Staff listed possible future uses: handling public records requests, searching old city records, reviewing building plans, creating renderings of proposed projects and watching for cyber threats, the Daily Journal reported.

A resident's suggestions

Half Moon Bay resident Evelyn de Souza, who said she has worked extensively in privacy, regulatory strategy and AI governance, called the draft a good first step. She suggested several additions:

  • A full inventory of the city's AI tools and disclosure of significant incidents.
  • Specific protections for high-impact uses like hiring and permitting.
  • A way for people to challenge errors made by AI.
  • A separation between the city's own monitoring and independent review.

The Oct. 20 item was a report and discussion, not a vote, so the policy has not been adopted yet. De Souza also asked the city to bring back a draft that shows where Half Moon Bay's version differs from the GovAI materials it was compared against, the Daily Journal reported.