
Portland City Council President Jamie Dunphy wants to hit pause on the city's rush into artificial intelligence, proposing a temporary moratorium on new AI contracts and any expansion of AI use across city government. The measure is headed to the full council for a vote on December 3, and it lands just weeks after records surfaced showing city technology staff and police quietly mapping out a much bigger AI rollout than the public had been told about.
Dunphy Says the City Isn't Ready Yet
Dunphy's proposal would freeze new AI contracts and block expanded use of existing tools, according to KOIN.com. If approved, the station reported, the pause would stay in place until the city puts privacy infrastructure and guardrails in place. “My goal is to protect public privacy, security, and civil rights,” Dunphy said, framing the pause as a chance to take time for critical thinking rather than rushing ahead.
Central to Dunphy's argument is that Portland's own Data and Privacy Office, created in March, isn't actually up and running yet. He said the office established in March is not yet operational and currently lacks the expertise or clear code guidelines needed to properly review technology contracts, per the same KOIN.com report. Without that office functioning, he argues, the city has no real mechanism to vet AI deals before they're signed.
A 365-Day Timeline Floated at Committee
The push for a pause didn't come out of nowhere. At an Oct. 29 City Council Finance and Governance Committee meeting, Dunphy said he planned to file an ordinance establishing a temporary 365-day moratorium on the city's procurement and implementation of AI, according to Street Roots. That timeline differs somewhat from how the pause has more recently been described in relation to the city standing up its privacy safeguards, and the dossier does not resolve which framing reflects the ordinance's final terms.
Street Roots' reporting is also what first revealed how far city departments had already gone in planning for broader AI use. The outlet found that the Portland Police Bureau, the Bureau of Technology Services and other city technology staff held at least eight meetings between May and September dedicated to using the Microsoft Copilot generative AI system. That's notable because, as the outlet reported, the city's April approval of Copilot had restricted its use to employees outside the Police Bureau in the first place.
Records Point to a Much Bigger Police Rollout
Despite that restriction, records from an August funding meeting indicated more than 1,000 users inside PPB could eventually use the tool if implemented, Street Roots reported. That figure raises the stakes considerably beyond the narrower rollout the city had publicly described. The City of Portland's own April announcement said Microsoft Copilot Chat Basic had been approved for employees outside the Police Bureau, alongside a new administrative rule establishing standards for how AI technologies are used across all bureaus, according to the City of Portland.
Portland residents who want to weigh in have a direct path to do so: the city says they can submit written or verbal testimony regarding the City's approach to AI through the city website, per KOIN.com's report. With the council vote set for December 3, that testimony window gives residents one more chance to shape the outcome before members decide whether to adopt Dunphy's pause.
Portland's Dilemma Mirrors a National Pattern
Portland's predicament reflects a broader tension playing out in local governments nationwide. A survey of public-sector leaders found that 63% of municipal respondents cite limited internal expertise or staffing as the biggest barrier to adopting or scaling AI, according to the International City/County Management Association. That same survey, which gathered responses from 109 public-sector leaders including 54 municipal officials across 26 states and the District of Columbia, found only 33% of respondents had a defined AI policy or governance framework in place, while another 30% were still developing one.
The trust gap is just as stark. A separate survey found that public sector AI use cases grew 50% from 2024 to 2026, yet 86% of public-sector organizations viewed their own AI use as transparent to the public compared with just 50% of residents who agreed, according to Smart Cities Dive. The same survey, conducted between April and May 2026, found 45% of public-sector leaders reported regularly using generative AI already.
Governance groups have been racing to catch up. The Center for Democracy and Technology analyzed public-facing AI policy documents from 21 cities and counties, including Washington County, Oregon, and identified five common trends, among them advancing accountability and human oversight in decisions that incorporate AI. At least 15 of those guidance documents explicitly called for ensuring AI tools comply with existing laws, the group found.
Other local governments have already pulled back on specific AI uses after running into trouble. The Ulster County Sheriff's Office in New York announced on October 22 that it would not move ahead with artificial intelligence-assisted report writing after a trial period, according to GovTech. That same day, Ulster County Executive Jen Metzger signed an executive order banning the use of AI surveillance technologies within her own administration. Nearby in Kingston, Common Council members raised concerns about data collection, privacy and potential AI errors as they weighed a proposed AI-integrated plan-review software costing $54,700 in startup costs, though the official overseeing that proposal said the reports would still be subject to review by staff.
For now, Portland's path forward hinges on what the council decides on December 3. Whether members adopt Dunphy's pause, modify it, or let the city's AI plans proceed as they currently stand remains an open question headed into that vote.









