Councilor Jon Link introduced a draft ordinance at the Oct. 22 City Council meeting that would put formal guardrails on how City Hall can use artificial intelligence, from banning AI-written police reports to giving residents a way to challenge consequential decisions made with AI help. The proposal was referred to the Legislative Matters Committee for a recommendation.
According to The Somerville Times, the Somerville City Council is fashioning an ordinance aimed at regulating parts of the city's AI usage, and the city could ultimately adopt rules establishing acceptable-use policies for AI technology and its physical infrastructure. The draft, formally the Oversight of Artificial Intelligence Systems ordinance, would establish municipal guardrails and a governance structure for how the city uses AI, covering AI systems generally, the handling of sensitive personally identifiable information, an annual AI usage report, transparency requirements and enforcement.
Councilor Jon Link, who introduced the measure, expressed concern about the potential harms AI could cause communities and emphasized the need to restrict certain uses outright. Link said the ordinance proposes banning AI decisions about eligibility for housing stability support, and the draft would also prohibit AI-written police reports and the use of biased AI models in decisions that could contribute to what he described as desperate outcomes. “AI should not decide whether an artist receives a grant, whether a community member goes to prison, or where a library is built,” Link said, per the same report.
What the Draft Ordinance Would Require
The proposed ordinance would require an annual report describing what AI tools the city uses, and City Council Vice President Wilfred Mbah said the measure would create a process for residents to challenge consequential decisions once it is adopted. Mbah said the goal is to establish clear, responsible guardrails for AI use in city government, arguing that AI can help government deliver services more efficiently while transparency and accountability must align with civil rights and privacy protections.
Mbah said residents should have the right to know when AI influences decisions affecting housing assistance, city services, benefits, employment or other important opportunities, and that low-risk administrative AI uses should be distinguished from higher-risk applications affecting residents' rights, with appropriate human oversight built in. He framed the broader aim as ensuring technology serves the public interest, protects civil rights and maintains human accountability, saying the framework is meant to let residents benefit from innovation while protecting them and maintaining public trust.
Not every councilor views the lines as easy to draw. Councilor Ben Wheeler said the ordinance's boundaries are difficult to define and that clarity is the guiding goal as the committee works through the draft. Councilor Lance Davis struck a more cautious note, saying organizations should be careful about the hazards and risks new technology can introduce, even as he said people should recognize the value and inevitability of widespread technology. Davis added that he looked forward to input from the community and the Wilson administration as the proposal moves forward.
What City Hall Is Actually Using Today
Separately, the city's own accounting of its current AI footprint turned out to be narrower than the ordinance debate might suggest. Deputy IT director David Slonina told the committee that Blue Voice, a police product that summarizes policies and advises officers, was the city's only, in his words, “true AI product,” as reported by Cambridge | Somerville Independent. Slonina said Blue Voice is trained only on Somerville policies and transmits no data, and that he knew of no city contracts or software products specifically built around AI tools such as ChatGPT, Gemini or OpenAI.
At the same time, Slonina acknowledged that AI shows up incidentally across city technology in ways that are easy to overlook, such as in sorting email, running traffic lights or preventing viruses, according to the outlet. That gap between a formal AI product list and the AI quietly embedded in everyday software is part of what the pending ordinance is meant to address. A separate resolution led by Link along with councilors JT Scott and Naima Sait also passed, calling for commonsense policies on AI-created media and AI use in staffing, per the same account.
How Somerville's Approach Compares
Somerville would not be the first city to formalize AI rules; it already ordained a Public Oversight of Surveillance Technology ordinance in 2019 requiring City Council approval before departments acquire new surveillance technology, along with annual reports on approved use, according to the City of Somerville ordinance text. On AI specifically, San Francisco amended its own AI inventory ordinance in committee in November 2024, requiring a public AI inventory and reporting, per San Francisco's Legistar records. The ordinance calls for departments to disclose an AI technology's purpose, accuracy, biases and limits.
Research from the Center for Democracy and Technology, which analyzed public-facing policy documents from 21 cities and counties, found five recurring themes in local AI governance: existing guidance, legal compliance, risk mitigation, transparency, and accountability paired with human oversight. The review also discussed legal compliance and risks related to AI use.
For now, Somerville's own draft remains with the Legislative Matters Committee, which will weigh in before any version goes to the full council for a vote.
