Board hears pitch for districtwide AI camera alerts; parents raise privacy, cost concerns

New Hanover County Schools trustees heard a presentation and public comment Tuesday on a proposed districtwide pilot to add AI-based object-detection to school camera systems, a plan backed by a state budget appropriation but met with strong questions from parents and staff.

The technology vendor Evident demonstrated software that runs inferencing at the camera or on local hardware to detect events such as a possible weapon, smoke or fire, slip-and-fall, perimeter intrusions, crowd panic and prolonged loitering. Evident said alerts would be routed to a mass-notification and incident-management platform called Raptor; the district would use its existing video management system to review footage after an alert.

Why it matters: the state legislature earmarked funding for school safety technology, and New Hanover County Schools would receive that money only if the board accepts a vendor and scope of work. Supporters say automated alerts could reduce response time for medical emergencies or trespassing; opponents say the systems raise long-term privacy, cost and mission concerns and worry about potential later activation of disfavored features.

Sean Hall, a presenter for Evident, said the system is not a facial-recognition product and that the AI models are "static" — trained in a lab and not learning from live deployments. "We're not doing any facial recognition. We're not tracking any information about students," Hall said. He also described the vendor's workflow: the video feed is analyzed, an alert is issued if the rules engine detects a defined event, and the vendor deletes the transient analysis data after processing. Hall added that some temporary buffering of video while inference runs — "seconds to a few minutes" — may occur before deletion.

District leaders and residents pressed Evident on deployment, data handling and cost. Superintendent Dr. Barnes said the technology was part of a multiyear effort and told the audience, "If our children are in imminent danger, or if our staff is in an emergency situation where there is imminent harm, I'm going to use every tool at our disposal to keep our kids safe." Board member Dr. Merrick said he would only back a contract if the state continued to fully fund the program and if facial recognition could not be activated: "I would only support this if the state continued to pay for it 100%."

Key technical and program details discussed at the meeting: - Scope and timeline: Evident proposed rolling deployment across all district schools with an initial delivery plan the company estimated at approximately three months, implemented in blocks of schools. - Edge inferencing and privacy: Evident said inferencing runs on camera-edge or local servers and that the vendor does not send raw video to its cloud; Raptor would receive only alert metadata. The company said permanent video storage would remain inside the district's existing VMS (video management system). - No facial recognition in current pilot: Evident confirmed facial recognition is not included in the proposed implementation; the company said the capability exists in its broader product line but would remain turned off unless the district and community later decided otherwise. - Pilot use cases: Evident defined five pilot use cases (weapons/weapon presence alerts, smoke/fire detection, slip-and-fall, perimeter/intrusion, and crowd movement/loitering). The district and vendor said additional capabilities exist in the product but would not be enabled without further approvals. - Data retention window: Hall acknowledged transient buffering of footage for analysis may last from "a few seconds to a few minutes" before deletion; the VMS would remain the authoritative archive for post-event review. - Vendor experience and K–12 track record: Evident said its AI has been deployed widely in airports, universities and other large venues but that it currently does not have K–12 deployments in the United States. - Procurement and funding: The district staff said the RFP was issued in October 2023 and that three vendors responded; a state budget amendment tied to a legislative appropriation determined vendors available to districts in that funding line. Meeting participants mentioned a state appropriation in the discussion of about $3.2 million for the program and that the Raptor subscription would be covered for the first two years by the grant; subsequent year-to-year subscription costs and longer-term maintenance were described as the board's future responsibility.

Parents and community speakers focused their questions on unintended uses, the potential for future activation of tracking features, accuracy and false-alarm rates, the district's existing camera quality, and ongoing operational costs. Child advocate Gail Major asked specifically whether facial recognition was included; Evident's presenters reiterated it was not included in the current proposal. Speakers also asked how administrators would reliably respond to push notifications while teaching and who would audit or independently evaluate the system during a pilot.

Some commenters asked whether the system would be used for student discipline or attendance enforcement. Evident representatives said the product is not designed to identify individual students for discipline; the company described perimeter alerts and other aggregate event detections but did not claim the system could identify named students leaving campus. The superintendent and several board members acknowledged those distinctions but said parents' concerns about surveillance and mission creep were legitimate.

Costs were a recurring topic. Meeting testimony referenced an implementation figure near $300,000 for the Evident software and hardware work and an annual Raptor subscription figure mentioned in the meeting (figures were disputed in the room). Board staff said the vendor response documents and RFP contain the detailed price schedule.

Votes at a glance: the meeting included several procedural board motions unrelated to the vendor decision. The board approved the meeting agenda (6–0), unanimously approved two successive 15-minute extensions of the forum (6–0 each), and adjourned by unanimous vote at the end of the session (6–0). No formal contract award or final board vote on the Evident/Raptor proposal occurred at this meeting.

Next steps: board members asked for additional community feedback and more detailed cost and scope clarifications. Several members said they wanted contract language that would keep facial-recognition features disabled unless the board and community explicitly approved them, and at least one member said he would only support a contract if the state continued to fully fund it beyond the grant period. The district asked attendees to submit comments by email and said staff would circulate the vendor's RFP response documents and pricing to board members prior to a later vote.