Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Education Ai Policy topic
No spam. Unsubscribe anytime.
Watchung Hills board tables draft AI policy after members question 25% detector threshold
Summary
Board members debated proposed policy 2365 on generative-AI use in classrooms, raising concerns about a 25% detection 'flag', student-data protections and the need for staff training. The board moved to table the policy pending clearer technical guidance and further committee review.
The Watchung Hills Regional High School District board spent a major portion of its June meeting debating a proposed generative-AI policy, policy 2365, before voting to table action pending clarification about a 25% detection threshold that would trigger human review.
The policy as presented would authorize district-wide AI use while leaving final discretion to individual teachers, introduce a standardized three-level AI-use scale for assignments, prohibit submission of fully AI-generated work, forbid inputting private student or staff information into public AI models, require citation of AI use when permitted, and list approved platforms such as Google Workspace Intelligence, Gemini, Microsoft Copilot, Canva and Adobe Creative Suite.
"We do not currently have an AI policy," said Liz, the district staff member presenting the draft, noting the plan was to bring the document forward for a first read and then a second read and vote before the start of the school year. She said the policy grew from a year-long professional learning community (PLC) that involved about 35 staff members and that the PLC reviewed templates from other districts.
Board members expressed a range of concerns. Several questioned whether commercially available AI detectors can reliably distinguish AI-authored text from student work and whether a numeric "25%" flag is a meaningful or stable trigger for review. Others urged the district to protect student data by limiting which platforms are permitted and to ensure human review and due process when a detector raises concerns.
One board member summarized the worry that detectors are imperfect and urged the policy to emphasize human review: if an automated tool flags a submission, that result should only trigger a teacher review and a student conference rather than automatic discipline.
District staff responded that the policy language already specifies a human-review step and that the first read is intended to allow technical clarifications before any final adoption. The board discussed options: approve a first read now and refine the definition of the trigger before the second read, delay the first read until the July meeting, or convene a focused education-committee meeting this summer to resolve outstanding issues.
After discussion, members voted to table further action on policy 2365 until the district can provide clearer technical information about the 25% threshold and related implementation details. Board leadership reiterated that a first-read approval would not make the policy effective; final adoption requires a second read and a vote.
The board asked staff to collect additional information over the summer — including the technical basis for the detector threshold, examples from peer districts and proposed training for teachers and students — and to route comments through Liz to the committee ahead of the next reading. The board also discussed offering orientation or coursework for students (for example, for incoming freshmen) on responsible AI use.
The policy debate was part of a broader set of policy updates the education committee presented to align district rules with the New Jersey Administrative Code; the board approved other routine policy changes on a first-read basis during the same meeting.
