Central York policy committee reviews draft AI policy; agrees to clarifications, plans further review

The Central York policy committee reviewed a draft artificial intelligence policy and agreed on several clarifications while asking staff to prepare revised language and return for one more short review before a planned first reading.

Committee members said the draft should keep a clause that the district "shall not be responsible for unauthorized charges or fees" as a notice to users and asked staff to place it in the general guidelines portion of the policy. Dr. Ucheff, who summarized legal guidance for the committee, said the clause primarily functions as notice and "ultimately any use of AI tools places the risk from that use between the student staff member and the AI tool and not the district." He advised that the language would not expose the district to liability in the scenarios covered by the draft.

The committee discussed several programmatic and compliance items: age-appropriate instruction on use of AI tools should be provided at all grade levels and integrated into existing digital-age and language-arts coursework (the district currently starts that curriculum in seventh grade), the director of technology must vet and maintain a whitelist of approved AI tools under FERPA protections, and the draft should explicitly reference the Children’s Internet Protection Act when addressing Internet filters. The draft policy requires that no personally identifiable student information be used with AI tools that are not preapproved by the director of technology or a designee.

On academic honesty, the committee retained a four-level rubric adapted from the North Carolina Department of Public Instruction that ranges from 0 (no AI use) to 4 (full AI use with human oversight). Members asked staff to add an explicit statement that if an assignment does not specify a permitted level, the default assumption will be 0 (no AI use). The committee also kept language allowing AI-detection tools as an investigative aid but prohibiting reliance on detection results "as the sole determination of academic integrity."

Members raised implementation concerns about detecting AI-generated work, noting available detection software is imperfect and that teachers will need practical tools such as lockdown browsers, revision-history checks, and authentic baseline assessments. The committee emphasized that discipline for misuse would follow the student code of conduct and that detection results should not be used alone to impose sanctions.

The policy draft also addresses operational uses of AI: the superintendent or designee must establish protocols to verify AI output used for budgeting, payroll, forecasting or HR analytics; final employment decisions (evaluations, promotions, hiring) must remain human decisions. The board will direct annual professional development on ethical and appropriate AI use for administrators, teachers and staff, with the committee favoring language that requires training "at least annually."

For monitoring and reporting, the draft lists the superintendent (or designee), building principals and the director of information technology as conduits for issues identified during periodic assessments of AI impact on student learning. The committee removed the word "generative" from policy descriptors to avoid narrow definitions and agreed to add a one-sentence explanation of CIPA in the compliance section rather than creating a definition list for every cited statute.

Committee members asked staff to draft the default-assumption sentence and other agreed edits, to prepare the student-handbook language that will accompany the policy, and to circulate revisions for a short final review in June. If no substantive changes are needed after that review, the committee aimed to present the policy for first reading on the June board agenda.