Technology

Schools experiment with staff training and traffic-light labeling to navigate classroom AI

Institutions like Cheshire Academy are focusing on general AI techniques and student reflection rather than prescribing specific software.

Updated Aug 24, 2026 · 15:23 UTC1 reports

The short version

Educators are grappling with increased workloads and confusion regarding generative AI, prompting some schools to offer general AI training rather than mandating specific tools.
Cheshire Academy in Connecticut has implemented a "traffic-light" assignment labeling system (red, yellow, and green) to define acceptable AI usage levels for students.
Teachers are leveraging both general chatbots and education-specific platforms like MagicSchool to draft lesson plans, worksheets, and grading rubrics, while piloting student-led councils to promote ethical use.

Cheshire Academy, a private school in Connecticut with approximately 400 students, trained its staff on general AI techniques, such as prompting and recognizing AI limitations, rather than prescribing specific tools.[MIT Technology Review]
The academy utilizes a traffic-light color system for assignments where green allows full AI use, yellow permits specific tools like spell-check while banning chatbots, and red bans AI entirely.[MIT Technology Review]
Some teachers design lessons that require students to actively evaluate AI outputs, such as having students review LLM-edited homework or peer-review AI-assisted assignments.[MIT Technology Review]
Educators are utilizing tools like ChatGPT, Perplexity, and MagicSchool—an education-specific AI platform that costs just under $100 per year for an individual paid plan—to generate administrative materials.[MIT Technology Review]

Some educators remain hesitant to use AI for generating student-facing materials or providing student feedback due to concerns regarding privacy, personalization, and accuracy.[MIT Technology Review]