OpenAI launches a classroom guide for teaching with ChatGPT

While schools worldwide are still debating whether students should be banned from using AI, OpenAI is taking the opposite route: handing teachers the playbook.

OpenAI just released an official guide for educators using ChatGPT in the classroom, featuring real-world teaching case studies, an FAQ covering bias and AI detector efficacy, and plug-and-play prompts developed with Wharton Interactive.

How educators are already using it

The guide highlights several educators around the world who have already embedded ChatGPT into their daily workflows.

At Old Dominion University, instructional technology professor Dr. Helen Crompton encourages graduate students to use ChatGPT for role-playing, having it act as an argumentative debate partner, a demanding recruiter, or a specific boss to uncover new angles.

In Spain, Universidade da Coruña professor Fran Bellas uses it to draft quizzes and syllabus ideas, making sure test questions are culturally relevant, accessible, and tailored to students' grade levels.

Meanwhile, Dr. Anthony Kaziboni at the University of Johannesburg uses the chatbot to help non-native English speakers polish academic writing and practice conversational English, leveling the playing field in academia.

In India, Geetha Venugopal at the American International School in Chennai teaches students not to take AI answers at face value, forcing them to verify outputs against primary sources to build critical thinking.

Structured prompts over one-shot answers

To help teachers hit the ground running, the guide includes GPT-4 prompts created by Ethan and Lilach Mollick from Wharton Interactive.

Instead of asking the model to generate a lesson plan in one shot, these prompts instruct ChatGPT to act as an instructional coach. It introduces itself, asks for the grade level, and pauses. It waits for the teacher's input before probing into prior student knowledge, learning goals, and potential misconceptions.

OpenAI also tackles the elephant in the room through its new FAQ, addressing model hallucinations, bias, and the questionable reliability of AI detection software.

Why it matters

We think this is a clever and necessary PR move by OpenAI to disarm institutional resistance in education.

By elevating respected academics rather than preaching top-down, OpenAI shifts the conversation from "AI is enabling cheating" to "AI is an assistant for overwhelmed teachers." Giving educators structured, interactive prompts prevents ChatGPT from spitting out generic lesson plans that nobody wants to use.

For teachers, it is worth testing immediately to draft practice questions or brainstorm role-play exercises. But the caveat remains clear: fact-checking and hallucination management will still fall entirely on the teacher's shoulders, meaning AI saves preparation time only if you are already an expert in what you are teaching.