I recently heard of two clever ways for teachers to incorporate generative AI, LLMs, and the like in classroom instruction — and I like them both.
The first came via Holden Thorpe, EIC of Science at AAAS, who described a classroom approach in which he tells students they can use ChatGPT all they want to prepare for a writing assignment — look things up, have it write passages, etc. However, the day of the writing, they are only allowed to bring in a single sheet of handwritten notes to class, containing no complete sentences, and from that they have to write the assigned essay.
This approach seems to intensify students’ involvement with relevant content, having to read, process, and distill the information numerous times prior to writing an essay based mainly on their actual knowledge.
The second example came via Charles Liu on the “Star Talk” podcast, where he described an assignment in which students had to formulate a question to enter into ChatGPT or a similar LLM, and then produce a report analyzing whether the answer was correct, incorrect, or somewhere in-between.
I’m sure there are myriad other approaches to clever uses of LLMs in classroom instruction, but I wanted to highlight these as a sign that, yet again, experienced teachers know how to use tools to drive content engagement for students, rather than letting tools drive them. And good teachers know how to teach students that mythologies — about technology, about anything — are to be analyzed and understood, not accepted uncritically.