Haverford forms AI working group; staff training and a traffic‑light protocol recommended for classroom use

Robert (Rob) Anderson, director of technology, and Meredith Hearn, technology integration coordinator, updated the committee on the district’s AI working group, its goals, training plans and next steps for staff and students.

What the working group is doing: Anderson said the district formed a roughly 30‑member working group after the administration adopted an administrative regulation to the acceptable use policy permitting staff engagement with AI. The group split into elementary and secondary teams and met regularly to develop guidance, drawing on external resources (AI for Education and Skills 21, Chester County/DIU workshops and multi‑district peer groups).

Training and teacher guidance: Hearn described a teacher training program called AI 101 that the district has begun delivering. Training covers basic definitions (what AI is and is not), district guardrails, and a “human‑in‑the‑loop” approach that requires human oversight at multiple stages: tool selection, input, output review and continuous auditing. Hearn said middle‑school staff received training in December; elementary sessions were underway and high‑school sessions were scheduled to follow.

Student usage framework: The working group proposed a traffic‑light protocol for classroom use. Red means no AI allowed for a given assignment; yellow allows AI for early‑stage assistance such as brainstorming or background research while preserving student‑authored work; green permits AI to assist in content development but requires that the final product be the student’s original work. Hearn said students would be required to submit a disclosure statement when an assignment permits AI, describing how AI was used.

Policy and implementation questions: Anderson and Hearn described a staged approach—staff may trial multiple tools now; the district will later consider narrowing to a small set of district‑endorsed tools after further evaluation. Anderson noted that current detectors for AI‑generated text are unreliable and can disproportionately flag English‑learners or neurodivergent students; he said the district is emphasizing assessment redesign (authentic, in‑class demonstrations of learning) and teacher familiarity with student work over sole reliance on detection software.

Concerns raised: Committee members and public commenters pressed the district on bias in AI outputs, representation of disabled students, environmental costs and human labor used in AI systems, and how teachers will safeguard students exposed to sensitive content. Anderson and Hearn said those topics are explicit parts of trainings, that the district is coordinating with county groups (DCIU, Chester County IU, DVCE) to get up‑to‑date resources, and that student lessons on AI ethics and environmental impact are being developed.

Timeline and next steps: Anderson said the working group will continue teacher training this winter and spring, evaluate tools and compile a technology plan. The district plans to present a broader technology plan to the board in March and to continue iterative review and auditing—Hearn emphasized the work is intended to be ongoing and responsive as tools evolve. Anderson added that student access to AI tools will be phased in only after staff training and appropriate safeguards.

Ending: The committee asked the administration to present the AI work to the full board; staff agreed to schedule a presentation and to return with more details about timelines, assessment strategies and measures to protect students and address equity concerns.