19473_Authority_Oct-opt

municipalauthorities.org │ 5 1 Diligence, owning what you then do with it. It is short, it costs nothing, and it gives an entire organization the same vocabulary. That last part is worth more than it sounds. [AI Fluency: Framework & Foundations, free as of August 9, 2026. https:// anthropic.skilljar.com/ai-fluency- framework-foundations ]. Then plan for the day it goes wrong. Write down what an AI incident looks like, who hears about it, and how quickly. Your staff need to know that flagging something odd is welcome rather than career-limiting. An authority that hears about small problems early is one that does not meet large ones by surprise. Proof that structure works Does any of this actually change anything? Our work at Las Virgenes Municipal Water District, in partnership with Sedna Consulting Group, gives us one. It is a California district, but the questions its staff brought into the room are the ones I hear from authorities here. Its board authorized an AI roadmap in November 2025 and received the results in open session on June 16, 2026. What stayed with me was not the technical assessment. It was what a single half-day session did to 17 staff from across the district. [www.lvmwd. gov/2026-06-16-board-regular-meeting] Every participant said it changed how they thought about AI priorities for their team. Seventy-one percent moved up at least one level in their own understanding of AI. Staff reporting a good or strong grasp of how it might apply to their work went from 41% to 82% in an afternoon. Set that against Pew's finding that 52% of American workers are worried about AI at work and most barely use it. Worry is the default position, and a policy document does not shift it. Putting people in a room with a structured way in does. [U.S. Workers Are More Worried Than Hopeful About Future AI Use in the Workplace, February 25, 2025, www.pewresearch. org/social-trends/2025/02/25/u- s-workers-are-more-worried-than- hopeful-about-future-ai-use-in-the- workplace/ ] One caution from that work. It is easy to get excited about the technology and end up experimenting for the sake of experimenting. In my experience, the organizations that get value start from a problem they already have and ask whether this helps. That discipline is the difference between a roadmap and a science project. Where to start Four things worth doing. Engage your board. Not to ask for a decision, but for their input and collaboration. That is how Las Virgenes began, and a board that helps shape the approach is one that backs it later. Write your hard rules before you choose a single use case. It is an afternoon's work, and it is what makes everything after it defensible. Share what you are learning in meetings you already hold. Ten minutes of a regular operations meeting, staff saying what they tried and what worked. They learn from each other, and you find out how this technology is really being used. Borrow from your neighbors. Other authorities are working through the same questions. So is the Commonwealth, whose own AI policy took effect in January and is published. Nobody needs to start from a blank page. [Pennsylvania Office of Administration, Artificial Intelligence Policy January 13, 2026, www.pa.gov/content/dam/copapwp- pagov/en/oa/documents/policies/ it-policies/artificial%20intelligence%20 policy.pdf] I opened by saying the only question is whether you can see it. Governance is how you see it. It will not guarantee that nothing goes wrong. It will tell you when something is going wrong, and what to do next. The greater risk is not that someone at your authority uses AI badly. It is doing nothing while it happens anyway. Get the governance right and the rest opens up. The authorities that get the most out of this will not be the boldest ones. They will be the ones who have implemented with intent, whose people know exactly what they are allowed to try and have been trained to get the best results. S 'ƵĂƌĚƌĂŝůƐ ĂƌƟĐůĞ ĐŽŶƟŶƵĞĚ ĨƌŽŵ ƉĂŐĞ ϭϳ͘ 5 1