 |
| Photo by Işıl via Pexels. |
Think-pair-share-- it's the easiest, most go-to way to get a conversation going in a room of people. It works well in rooms that don't have flexible seating for gathering into larger groups, it's a low amount of investment in setup and explanation, and it gives literally everyone in the room a chance to talk (unless one half of the pair is either real shameless or charmingly enthusiastic). And if done well, it's an "activity" that makes such intuitive sense it doesn't have to feel like an "activity."
If you haven't done think-pair-share before as a leader or participant, this description of the activity and FAQ about it is a nice introduction.
Normally, think-pair-share is conducted by humans (for NOW), but an AI chatbot can be one half of your pair or a supplement to a human pair. This was one of the first uses of ChatGPT I wondered about when it became a popular subject of conversation, and I wasn't alone! Several pieces offer ideas about incorporating AI into your Think-Pair-Share time (for example, it's one of Ditch that Textbook's 20 Ways to Use [Chatbots and Artificial Intelligence] as a Tool for Teaching and Learning). Most treatments of this idea I've seen are riffing off of this widely shared tweet by Sarah Dillard.
[An aside: one of the really funny things to me about the surge in conversations about AI in teaching is how often I now see people talking, writing about, and sharing tweets as though they're academic articles in circles where that might never have been the case before. When something is very new, many standard expectations of what kinds of things are acceptable or useful to cite shift; I imagine this is less wondrous to folks doing more present-focused work, who have frequently encountered the wonderful world of citing social media before, than to those of us who spent seven years citing while fighting microfilm-induced nausea.]
For some real zaniness, and for more insight into the tools themselves, you could have students fire up two chatbots and give them instructions, then feed their responses to one another. This can take careful prompting that will depend on the topic of discussion; it also may shed some light on the boundaries of the chatbots-- it's likely that their conversation will become a bit circular, as they tend to declare everything up front rather than have an evolving conversation (similar to some of the worst human small group activities I've been part of, really).
For more complex use of a Think-Pair-Share type framework, I love this set of options from Acadly, which suggest alternatives to insert into the process like think-write-pair-share, which encourages fuller consideration of the issue before the pair stage, and think-vote-pair-vote-share, which would work well for a question on which minds are likely to change after some collaboration or conversation on the topic. Generative AI could easily be incorporated into these steps to provide some real value; for example, comparing the written thoughts in think-write-pair-share with how ChatGPT might respond to the prompt, or asking the chatbot for reasons why someone might disagree with one's original vote in the think-vote-pair-vote-share framework.
