Apex IT staff outline cautious AI posture, pilot uses and privacy safeguards

Tom McInnis, who leads the town’s IT security and innovations division, told the council the town’s posture is to treat artificial intelligence as a tool that can improve efficiency but still requires human judgment and security safeguards.

McInnis explained the difference between predictive AI and generative AI and warned about risks such as hallucinations and bias: he recounted a recent example in which an AI‑generated legal brief included a non‑existent case, saying that in that incident "an AI told me it exists." He described current pilots that use predictive AI for parking occupancy heat maps, traffic and mobility modeling, pavement condition analysis with a vendor named Raven, crowd counting at events, and a chatbot pilot (Polymorphic) to support Ask Apex call‑center functions.

Policy and privacy: McInnis said the town’s existing IT policy governs third‑party data sharing and that the pilots avoid storing sensitive criminal‑justice or personal data in third‑party systems at this stage. He recommended a coordinated approach to avoid 'shadow AI' (unmanaged departmental usage) and proposed an innovation lab or working group—mixing staff and residents—to vet pilots and set priorities. Councilors emphasized resident privacy and small‑town character, and asked staff to prioritize use cases that produce operational efficiencies and preserve human customer service.

Next steps: staff asked for direction on priorities and said they would return with more detailed proposals and examples of in‑house and pilot tools that could be scaled; no formal policy was adopted at the meeting.