Gloucester County staff outline AI use and Placer location analytics; supervisors press for privacy, cost and contract details

Gloucester County staff presented the county’s current and planned uses of artificial intelligence (AI) during the Board of Supervisors meeting on Feb. 4, 2025, outlining both general AI tools such as ChatGPT and specific services the county is piloting, including Placer, a location-analytics vendor that aggregates mobile-device movement data to estimate visitors and patterns.

The presentation, led by county information-technology staff and department directors, described safeguards the county is using and proposing — onboarding and training, monthly review meetings, and a policy approved by the county attorney — while staff acknowledged limits and risks including hallucinations, biased outputs and the presence of third-party data collectors. Supervisors repeatedly asked how Placer obtains data, whether children or sensitive sites can be identified, how accurate the counts are, and how much the subscription costs.

Why it matters: county staff said the tools provide faster, broader data than legacy methods and help with event planning, tourism marketing, grants and traffic planning. Critics said the data collection is largely out of residents’ control and pressed for transparency on data sources, partner lists, contract language and whether the county should continue the subscription.

Britton Hogg, a software analyst, gave the technical overview, saying, “AI is already here whether we want it or not,” and described general AI services county staff have experimented with, including ChatGPT, OpenAI’s Operator and Deep Research, Microsoft Copilot and Google’s Gemini. Derek Beach, the county’s chief information officer, urged “responsible” adoption and said membership in a government AI coalition helped staff craft rules and training.

Placer’s product, described in the presentation, combines aggregated mobile-device data and “differential privacy” routines and is sold to governments and businesses for use cases such as estimating event attendance, drawing trade-area ZIP-code maps for economic development and informing park management. Matt Barber, director of parks, recreation and tourism, said Placer allowed staff to estimate event counts and visitor behavior: “What we can do with Placer is draw a diagram around what we consider the event site, and then we get the accurate counts for in that area,” he said, citing staff examples such as counts staff reported for the Daffodil Festival and other events.

Economic development staff said the tool supplies trade-area ZIP-code breakdowns and retailer-visit estimates useful to small businesses and recruiters. Sherry Spring, director of economic development, presented 12-month visit totals for a local shopping center and specific retailers drawn from Placer data, stating numbers in the presentation (for example, a high number of visits reported for a large retailer), and emphasizing the tool’s usefulness when writing grants or marketing the county.

In response to board and resident questions, staff said Placer does not give identification data for individuals and that the vendor requires aggregation thresholds (staff said at least 50 devices must be present in an area before a data point is produced). Staff also said Placer’s partners strip personal identifiers before aggregation and that the vendor prohibits knowingly supplying data on minors or on certain sensitive location categories, though board members pushed staff to provide the vendor partner list and the contract clause language that define the limits.

Supervisor Cristo said residents lack a practical opt-out from Placer’s underlying data collection and asked whether the county’s purchase makes a material difference, saying, “I don't have that option. I either have to turn my phone off and not be counted or I don't have an option to not use Placer as a citizen and an end user.” Staff replied that the county did not provide Placer with individual device data and that Placer acquires data from sources such as mobile apps and data brokers.

Accuracy and cost were repeated concerns. A supervisor asked whether Placer figures are tested against in-person counts; staff said they have discussed validation with the vendor and that the vendor compares some outputs to store sales and other benchmarks but that independent, public validation was limited. On cost, staff gave a first-year subscription figure described in the meeting as roughly $20,000; board members cited other cost numbers during the discussion, including travel and membership costs for a government AI group, and asked for a written cost breakdown before the board decides on continuing the subscription.

Board members and members of the public asked several specific follow-ups: they requested a copy of the county AI policy (staff said the county attorney approved it and staff would distribute it), a copy of the Placer contract and vendor partner list, documentation showing the methodology for device-aggregation and the 50-device threshold, a cost-benefit analysis tied to concrete use cases (for example, tourism grant success or avoided survey costs), and the possibility of a Placer representative presenting to the board. Staff said they would provide the requested materials and that, if the board instructs them to end the contract, they will do so.

The presentation and ensuing debate left the board without an immediate policy change. Staff said they will supply the policy and vendor materials and that they are prepared to end the Placer contract if the board directs it.

Ending: Board members offered two paths at the close of the discussion: some said they wanted more information and a written packet before deciding, while others reiterated privacy and opt-out concerns and asked for quick access to contracts and policy documents so the board and public can judge whether the county’s use of location-based analytics is acceptable.

Votes at a glance: The AI discussion produced no formal vote to authorize or terminate a contract during the meeting; staff indicated the current Placer subscription is a pilot-year subscription and that continuation would be considered after the board reviews the requested materials.