AI in Schools: New Approaches for Educators and Institutions

As AI tools rapidly infiltrate the educational landscape, schools and teachers continue to face pressure to establish clear policies. Meanwhile, pioneering institutions like Connecticut's Cheshire Academy are adopting a balanced model that relies on general prompt training and practical applications rather than mandating specific software. Although organizations such as UNESCO and OpenAI promote AI in education, many teachers still experience uncertainty regarding its classroom integration.

Cheshire Academy's Blended Education Model and Applications

With a student body of around 400, Cheshire Academy chooses to provide staff training on general AI techniques under advisory guidance, rather than forcing instructors to use a specific tool. Teachers and technology coordinators at the school blend general-purpose chatbots like ChatGPT and Perplexity with platforms specifically developed for educators, such as MagicSchool.

While the boundaries of this technology, its potential to generate biased responses, and its error margins are strongly emphasized during the educational process, teachers primarily leverage AI to assist with lesson planning and the creation of grading rubrics.

Critical Thinking and AI Practices in the Classroom

Some teachers, whether cautious about or experienced in AI technologies, position artificial intelligence as a tool to develop students' critical thinking skills rather than using it directly to produce lesson materials. For instance, in French classes, students are asked to have a language model edit their homework, and then examine the model's corrections to determine which ones are accurate and which ones undermine their own personal expression. Such methods enable students to experience firsthand the limits and blind spots of AI in language learning.

Sectoral Reflections and Points to Consider

As the role of AI expands in the education sector, a fine line exists between its potential to reduce teachers' workloads and concerns over quality, personalization, and data privacy. In particular, question marks surrounding student data privacy and feedback quality rank among the biggest obstacles preventing teachers from fully integrating AI into student assessment processes. Rather than imposing one-size-fits-all software, institutions are increasingly focusing on flexible, conscious usage models that teachers can filter through their own pedagogical judgment.

*This report was prepared based on data published by MIT Tech Review — AI.