District staff outline AI policy 5840, urge guided use and staff training before formal vote

District technology staff and board leaders discussed an artificial intelligence framework and a draft Policy 5840 at the board work session, describing a recommended approach that permits supervised use while building training, content filtering and approval workflows.

Ray Ranieri, who led the technology committee presentation, summarized several vendor and higher-education engagements and described two broad policy paths the committee considered: a highly restrictive ban on student AI use (dismissed), and a managed-use policy that allows directed, supervised AI applications coupled with training, approval and content controls. Ranieri said the recommended path emphasizes planned classroom use, staff training in “prompting” and the deployment of embedded, task-specific tools rather than an unrestricted, open-chat model.

Presenters listed specific risks and operational controls. They warned about so-called ‘‘hallucinations’’—instances where generative models produce inaccurate or fabricated content—and said predictive analytics can compound errors if not supervised. The technology committee described protections the district plans to use, including content-filtering software, an app-approval process, contractual/vendor reviews and use of Microsoft Copilot (bundled with the district’s Microsoft environment) for staff productivity tasks under supervised deployment. The district said it will pilot guarded or embedded AI features (for example, text-to-speech or task-specific assistants) rather than open generative-chat tools for student work unless a supervised classroom protocol is in place.

Staff reported the committee had convened multiple conversations with regional partners including BOCES, Niagara University and other districts. Those partners presented a range of approaches at recent meetings: BOCES emphasized a guarded, assistive stance; university partners outlined “moderate” classroom use and professional development; another district described a more open embrace of AI under teacher direction. Ranieri and the presenters said those conversations informed the draft policy language and the district’s proposed training program for teachers and administrators.

District leaders asked the board to expect Policy 5840 on the agenda at the next voting meeting and recommended tabling the policy after initial approval so the public, staff and families can review the text and provide feedback. The recommendation, as presented, is to approve a policy framework on paper, then keep the item open for about a month for comment and refinement before finalizing implementation details.

Board members asked operational questions about classroom detection, staff development and the means to ensure students learn underlying skills rather than only using AI to generate outputs. Presenters emphasized two staff-development strategies: training teachers to “know their students” so suspicious submissions can be probed, and training teachers in progressive prompting techniques that reveal depth of student understanding. Ranieri said the technology committee will continue iterative updates to the policy as tools evolve.

Ending

The board was told Policy 5840 will appear on the next meeting’s agenda; staff recommended tabling the measure after initial approval for a public review period. The district will continue pilots, training and vendor licensing steps while posting the policy text for community comment.