Boyertown Area SD staff recommend district AI policy that limits student tools to district‑approved resources

Boyertown Area School District officials presented a draft policy on use of generative artificial intelligence in schools and asked the board to consider a multi-step adoption process that would include annual review, an administrative regulation for tool approval, and limits on student access to district‑authorized tools.

The discussion centered on a draft based on PSBA policy 8.15.1 brought by Stephanie Landis, Director of the Office of Teaching and Learning, and Dave Feit, Director of Technology. Landis said the policy would require that “generative AI tools and resources used in district schools and programs shall be evaluated and authorized on an ongoing basis,” and listed evaluation factors including age appropriateness, bias, accessibility and data security.

Landis and Feit described existing protections already in place: district single sign‑on through the Clever platform, a Google Apps agreement limiting shared student data under COPPA and FERPA, routine phishing/security training, and professional ethics training for staff. Feit said the district would “contain the data that the app or the application or the website uses through our Clever instance,” so students do not create accounts or provide personal data.

Why it matters: the board heard that about 6,500 students could be affected and that the district needs a policy flexible enough to adapt to rapidly changing tools but stable enough to avoid constant rewrites. Landis recommended pairing the policy with an administrative regulation (AR) so teachers can propose new AI resources for vetting without asking the board to rewrite policy whenever a new tool appears.

Key recommendations and clarifications presented - Only district‑authorized generative AI tools would be permitted on district computers and in district programs for students; teachers would have greater flexibility for classroom use. - The policy would require AI literacy training for staff, students and IT personnel; Landis said that professional development and ethical training would be embedded in existing staff training cycles. - The draft permits the use of AI‑detection tools “as an aid” but prohibits relying on detection results as the sole determination of academic integrity; Landis and Feit recommended against purchasing AI‑detection products now because they are not sufficiently accurate and are the subject of ongoing litigation. - The policy as written originally called for prior informed parental consent before student use; Landis recommended replacing that requirement with an annual district notification to families given the practical difficulty of obtaining explicit opt‑ins from thousands of households. - The board was advised to decide which staff should evaluate and authorize AI resources; administration recommended the director of IT and the director of OTL be the co‑owners of that vetting process.

Board and public questions focused on scope, cost and enforcement. Board member Michael Zawada warned the board to “not just sit on it every year” and urged frequent reviews, asking whether the work should be a permanent committee or cabinet position given the scale of the change. Landis and Feit said some proposed student tools being reviewed would likely carry licensing costs; they referenced East Penn School District’s reported ballpark licensing estimate of about $7 per student per year and said exact figures were “not specified” for Boyertown.

Several board members raised academic integrity and virtual‑program enforcement concerns. Landis described classroom practices to detect improper use, including scaffolding writing assignments, conferencing between teacher and student, and teacher‑initiated follow‑up if submitted work differs markedly from a student’s usual output. For students in BASD virtual programs, Landis said those students remain subject to the district’s acceptable use and discipline policies, and teachers would investigate suspected misuse using the same steps as for in‑person students.

Next steps and timeline Landis said the administration will revise the PSBA draft to reflect the board’s input, return a revised draft to the Committee of the Whole in March and — if the board wishes — bring it as a first read at a legislative meeting, with a second read and potential adoption targeted for April–May so policy could be in place before the 2025–26 school year. The presenters repeatedly emphasized that the district would pair policy with an AR to allow staff to submit specific tools for vetting.

What the presenters declined or warned against Feit and Landis advised that the district not purchase AI‑detector software at present because of accuracy and pending legal questions. They also cautioned that students' off‑campus conduct involving AI would be treated consistent with existing student discipline policy to the extent that policy applies.

The board left the meeting with direction to the administration to return with a revised PSBA 8.15.1 draft and recommendations for implementation and family notifications.