Thanks, Mike. I think this will be a very valuable tool for school librarians working with students. I will be sharing at my next training session on Tuesday.
Thanks, this is very exciting! I've just been playing around with it, and I can see how valuable this might be for my students, but there's some inconsistency in whether it really follows the Toulmin structure. (These bots are squirrely things!) I wonder if you'd be willing to share the prompt or any tips on the prompt? I would love to play around with it further, maybe on PlayLab.ai where I have some bots. It would be great to collaborate on iterations of this...I'd love to do a version that goes with some of the ancillary materials for my textbook, but my textbook teaches the Toulmin method in slightly different way (I talk about assumptions, not warrants, and don't teach backing as it seems too similar to evidence--and other adaptations I developed after years of correcting students on vocabulary).
This is really cool, Mike. I'm gonna try it with students soon and report back. I like that it asks questions but I wonder if I like that it also answers many of the questions. In the sense that in a learning environment I want them to do this kind of asking for answering. But it does model it quite well. I tried it with something in Arabic and it crashed at first but in the end did a very good, nuanced job. I can see how this could augment or replace a site like snopes or hoaxbusters
Oh and it could be cleaned up a lot, and made to work better on Gemini too. I don't know why but lately it's been garbage on chatgpt. If you or your students do any helpful edits let me know.
Thanks, Mike. I think this will be a very valuable tool for school librarians working with students. I will be sharing at my next training session on Tuesday.
Thanks, this is very exciting! I've just been playing around with it, and I can see how valuable this might be for my students, but there's some inconsistency in whether it really follows the Toulmin structure. (These bots are squirrely things!) I wonder if you'd be willing to share the prompt or any tips on the prompt? I would love to play around with it further, maybe on PlayLab.ai where I have some bots. It would be great to collaborate on iterations of this...I'd love to do a version that goes with some of the ancillary materials for my textbook, but my textbook teaches the Toulmin method in slightly different way (I talk about assumptions, not warrants, and don't teach backing as it seems too similar to evidence--and other adaptations I developed after years of correcting students on vocabulary).
Can you message me your email on LinkedIn and we’ll get started ;)
Of course!
I have taught the Toulmin analysis methods to students for nearly 30 years, so this resource is very exciting!
Two questions?
(1) How did you train the GPT to identify implicit warrants?
(2) Have you given any thought into the possibility of a "Toulmin coach" GPT that could aid students in working through their own argument analysis?
Could you edit this to include a link to the video? For privacy reasons I have YouTube embedding disabled through my ad blocker.
Sure. I didn't realize it didn't do a smart default for that
This is really cool, Mike. I'm gonna try it with students soon and report back. I like that it asks questions but I wonder if I like that it also answers many of the questions. In the sense that in a learning environment I want them to do this kind of asking for answering. But it does model it quite well. I tried it with something in Arabic and it crashed at first but in the end did a very good, nuanced job. I can see how this could augment or replace a site like snopes or hoaxbusters
Oh and it could be cleaned up a lot, and made to work better on Gemini too. I don't know why but lately it's been garbage on chatgpt. If you or your students do any helpful edits let me know.