Mercer Island schools outline AI values, classroom controls and staff training

Mercer Island School District staff on Tuesday presented a framework for classroom use of artificial intelligence that centers on five stated values — transparency, safety and accountability, professional development, inclusive and equitable settings, and academic integrity. The presentation reviewed tools already in teacher use, safeguards the district plans to require, and how teachers are training students to use AI as a learning aid rather than a shortcut.

The district's technology lead told the board the work began about two years ago and has included guidance from the Office of Superintendent of Public Instruction (OSPI), staff, students and families. The presentation noted that some AI tools have age or privacy restrictions and that the district is focusing on tools that can be used across grade levels while protecting student data.

Administrators said the district has updated existing technology and harassment policies rather than writing a separate AI policy, following OSPI advice. Staff told the board they have added examples of inappropriate AI use to the district Responsible Use Policy and to student handbooks, and have begun sending separate notices to families at the start of the year explaining privacy rights, translation and opt-out options.

Staff described classroom controls used in a platform called Magic School (also mentioned as Magic Student), where students can access AI only through teacher-created “rooms” and unique room codes. Teachers can see every student prompt and the AI responses inside that room, a capability staff said gives teachers real-time oversight and a way to judge if students are using AI thoughtfully or copying content. District analytics presented by staff showed roughly 448 unique staff users of Magic School, more than 16,000 AI generations attributed to staff use, and nearly 9,000 student sessions across the year; staff said the tool’s top student uses were practice quizzes, writing feedback and research summarization.

The presentation also covered professional development: the district runs an annual Instructional Tech Day with about 50 workshops led largely by teachers; it operates technology learning cycles where coaches work shoulder-to-shoulder with each teacher on implementing tools; and it distributes brief “tech tip” messages throughout the year. Staff said they have additional supports for classified employees, including two half days of in-person training.

District staff acknowledged ethical and technical concerns raised in the discussion, including the environmental impact of large-scale model generation; potential bias in model outputs; copyright and authorship questions for text and images generated by AI; student privacy under federal rules; the limits of AI-detection tools for plagiarism; and the need to avoid overreliance by students. Staff said some of these topics are already being taught in project-based classes (for example, a Crest project that examined AI ethics and environmental impacts) and that the district has updated its citation guidance to require students to cite AI use when it is incorporated into work.

Board members and student representatives asked for follow-up demonstrations and study sessions. Several board members said they want the district to track federal and state initiatives and grant opportunities, and to study whether some AI uses could qualify as “high-impact tutoring” or support math interventions. Administrators said the district will continue to evaluate tools against FERPA and CIPA requirements and will push more communication and demonstrations to the board and community.

The presentation materials, administrators said, include links to federal guidance and the district’s AI values and classroom examples on the district website.