Schools test practical rules for classroom AI

Cheshire Academy, a private boarding and day school in Connecticut with about 400 students in grades 9 through 12, has taken a flexible approach to AI in education. Administrators do not require instructors to use AI, but the school’s librarian and technology coordinator George Aiello says the “vast majority” use it in some way, including ChatGPT, Perplexity and MagicSchool.

The school trained staff on general AI techniques rather than mandating specific products. Teachers learned prompt-writing practices as well as the technology’s limits, including the risk of inaccurate or biased output. Many now use generative AI to prepare lessons, create grading rubrics and develop classroom materials, while concerns about quality, personalization and privacy have slowed adoption for student feedback.

French teacher Miriam Przybyla-Baum has focused on helping students evaluate AI rather than relying on it. Her assignments ask students to review LLM edits to their homework and judge whether changes are correct or weaken their voice. The school has also adopted a traffic-light labeling system for assignments: green allows AI, red bans it and yellow permits only specified tools.

Cheshire Academy is also piloting a “Student AI Council,” where students create media and lead discussions about responsible AI use. The broader lesson is to meet students where they are, model careful use and treat AI output as material to refine rather than final work.