Latest / Azadeh and Guy on AI / What Students Should See Us Doing With AI?

Transcript

What Students Should See Us Doing With AI? · Azadeh and Guy on AI

• 0:00
  SPEAKER_02: Hi, and welcome to another episode of Azade and Guy on AI.
• 0:06
  SPEAKER_02: And we have Nick with us for a second round.
• 0:08
  SPEAKER_02: We were so excited about the conversation in our previous episode that we had to do another one.
• 0:15
  SPEAKER_00: Yeah.
• 0:15
  SPEAKER_02: Absolutely had to.
• 0:17
  SPEAKER_00: And every time we make him laugh, then we introduce the podcast.
• 0:22
  SPEAKER_01: Well, it's really more like I was here and they were like, we should do a second episode.
• 0:26
  SPEAKER_01: And I was like, we still have questions.
• 0:28
  SPEAKER_01: Sure.
• 0:28
  SPEAKER_01: Let's go.
• 0:29
  SPEAKER_01: Convenient sampling, people.
• 0:33
  SPEAKER_02: Anyway.
• 0:34
  SPEAKER_02: All right.
• 0:36
  SPEAKER_00: So I'm going to ask the first question for this second take.
• 0:40
  SPEAKER_00: Most of our pre-service teachers will learn more from how we use AI
• 0:44
  SPEAKER_00: in our own courses than from anything different.
• 0:48
  SPEAKER_00: So what do you want them to see you doing and what do you hope they never see you doing?
• 0:59
  SPEAKER_01: So I think you're dead on with this question, right?
• 1:04
  SPEAKER_01: Like starting from Lordi's construction of how pre-service teachers learn
• 1:12
  SPEAKER_01: to be teachers, that notion of the
• 1:17
  SPEAKER_01: First thing you know is the educational processes that you yourself have experienced.
• 1:24
  SPEAKER_01: And so for myself,
• 1:28
  SPEAKER_01: in my reading coursework,
• 1:30
  SPEAKER_01: a lot of what I end up doing is undoing not great instruction, right, in the past.
• 1:41
  SPEAKER_01: And part of that undoing is a lot of modeling.
• 1:46
  SPEAKER_01: And so thinking about AI and how AI can get used in education is I do a lot of narrating.
• 1:58
  SPEAKER_01: I do a lot of pedagogical narration.
• 2:01
  SPEAKER_01: I do a lot of being very clear and transparent, which feels a little radical at times.
• 2:10
  SPEAKER_02: Mm-hmm.
• 2:10
  SPEAKER_01: Right now,
• 2:11
  SPEAKER_01: because it's,
• 2:12
  SPEAKER_01: oh, did you use AI because you can't think about this yourself on your own?
• 2:15
  SPEAKER_01: It's like, no,
• 2:16
  SPEAKER_01: it's because I am a single person with three dogs and bills to pay and a kitchen
• 2:22
  SPEAKER_01: to clean and a car to take into the garage.
• 2:25
  SPEAKER_01: Like, if it's going to buy me time or allow me to do something better,
• 2:31
  SPEAKER_01: if it's going to take my core skill and elevate it, why wouldn't I do it?
• 2:37
  SPEAKER_00: Yeah.
• 2:39
  SPEAKER_01: Right?
• 2:39
  SPEAKER_01: Like...
• 2:40
  SPEAKER_01: Keeping in mind that, yep, it does take environmental resources to make that happen.
• 2:47
  SPEAKER_01: However, I'm not Googling what's my best sweater for fall 2026, right?
• 2:54
  SPEAKER_01: I'm using it to do actual work that improves the lives of my students.
• 2:57
  SPEAKER_01: And so that's how I justify that.
• 3:00
  SPEAKER_01: So I try very hard to
• 3:06
  SPEAKER_01: I don't try.
• 3:06
  SPEAKER_01: I do.
• 3:08
  SPEAKER_01: I narrate my AI use out loud.
• 3:11
  SPEAKER_01: I am really clear with students about how I'm using it.
• 3:16
  SPEAKER_01: I'm using it to help me hold the memory of readings from four weeks ago
• 3:22
  SPEAKER_01: so I can put them in conversation with two days reading.
• 3:28
  SPEAKER_01: That is something that my adult brain Isn't always good at because I'm thinking
• 3:37
  SPEAKER_01: about, oh,
• 3:38
  SPEAKER_01: I have to go staple chicken wire to my fence
• 3:41
  SPEAKER_01: because my neighbor's dog is really aggressive.
• 3:43
  SPEAKER_01: And Rhoda is now, like, Rhoda, my dog,
• 3:47
  SPEAKER_01: is now all bloodied up because she fought the fence and the fence won, right?
• 3:51
  SPEAKER_01: So, like, it frees up brain space, right?
• 3:56
  SPEAKER_01: Like, it allows me to do some of that.
• 3:58
  SPEAKER_01: So I try to make sure that...
• 4:03
  SPEAKER_01: I provide those clear parameters around this is how I'm using it, this is how I'm not.
• 4:09
  SPEAKER_01: And it's sometimes required, and I do that for them too.
• 4:13
  SPEAKER_01: Like, here's how I want you to use AI for this assignment.
• 4:17
  SPEAKER_01: Like this one, you all have to use AI.
• 4:20
  SPEAKER_01: Because you're going to encounter it whether you love it or not.
• 4:25
  SPEAKER_01: And so it's literally on their assignment sheet.
• 4:29
  SPEAKER_01: Like, this is how you use AI, this is how you don't.
• 4:32
  SPEAKER_01: If there's specific prompts I want them to use, I give those to them.
• 4:37
  SPEAKER_01: And I really try to make sure that they have a good sense of what is a good prompt
• 4:45
  SPEAKER_01: for chat or Claude, whatever it is.
• 4:47
  SPEAKER_01: And then I also...
• 4:55
  SPEAKER_01: Like narrate my failures, which there are a ton of them, right?
• 5:00
  SPEAKER_01: Like, I think that given the environment in which AI is happening right now,
• 5:12
  SPEAKER_01: one, there's not a lot of space for failure, one, just writ large.
• 5:16
  SPEAKER_01: And then the notion of like AI is cheating at life somehow, right?
• 5:24
  SPEAKER_01: But it's not.
• 5:24
  SPEAKER_01: It's just allowing me to do better things.
• 5:26
  SPEAKER_01: So, for example,
• 5:29
  SPEAKER_01: I use exams,
• 5:32
  SPEAKER_01: I use quizzes in my course,
• 5:33
  SPEAKER_01: multiple choice, great for retrieval practice, really helps students get into it.
• 5:39
  SPEAKER_01: And I've been using AI this semester to be much more methodical about my item construction.
• 5:48
  SPEAKER_01: Where I've done kind of a,
• 5:50
  SPEAKER_01: this summer I did kind of a synthesis of the multiple choice question genre.
• 5:56
  SPEAKER_01: What do we know from research?
• 5:59
  SPEAKER_01: What makes up a good one?
• 6:00
  SPEAKER_01: Et cetera, et cetera, et cetera.
• 6:02
  SPEAKER_01: And have built a skill that takes my previous questions and which were,
• 6:11
  SPEAKER_01: I mean, okay, but it's just me.
• 6:14
  SPEAKER_01: You know,
• 6:15
  SPEAKER_01: single doubting it for all these students trying
• 6:17
  SPEAKER_01: to create 40 multiple choice questions for, you know, every reading.
• 6:23
  SPEAKER_01: And it literally, Claude just worked for like an hour and revised all of them.
• 6:32
  SPEAKER_01: And one of the rules that I had for Claude as they were working on
• 6:37
  SPEAKER_01: that project was you're not referencing, you're referencing concepts.
• 6:41
  SPEAKER_01: You're not referencing concepts.
• 6:43
  SPEAKER_01: Specific articles, right?
• 6:45
  SPEAKER_01: So, like,
• 6:47
  SPEAKER_01: it would sometimes come up with questions like,
• 6:52
  SPEAKER_01: according to this article,
• 6:54
  SPEAKER_01: author here on insert page number here, there's this insert table number here.
• 7:00
  SPEAKER_01: What does that table mean?
• 7:01
  SPEAKER_01: Not a great question.
• 7:03
  SPEAKER_01: Right?
• 7:04
  SPEAKER_01: Not a great question.
• 7:05
  SPEAKER_01: And even if 9 out of 10 were great, I still had to do some pruning.
• 7:12
  SPEAKER_01: And...
• 7:13
  SPEAKER_01: Even if I'm doing pruning, sometimes I'd miss one.
• 7:17
  SPEAKER_01: Or I'd miss two.
• 7:18
  SPEAKER_01: Or I'd miss three.
• 7:19
  SPEAKER_01: No, I haven't missed three.
• 7:21
  SPEAKER_01: Normally, it's just like I'm down to two.
• 7:24
  SPEAKER_01: But then talking through, this is my bad.
• 7:30
  SPEAKER_01: I went through, I knocked out the rest of these questions, but I missed one.
• 7:36
  SPEAKER_01: And here's why this is an unfair question.
• 7:38
  SPEAKER_01: Let's talk about this as a distractor and why I would not...
• 7:43
  SPEAKER_01: Why this would not be something that I would keep here, right?
• 7:47
  SPEAKER_01: And
• 7:50
  SPEAKER_01: that's been really helpful for me in terms of keeping me honest.
• 7:56
  SPEAKER_01: And I'm hoping that it's also making a clear narrative for students who might think
• 8:05
  SPEAKER_01: that the product that AI is going to give you is going to be perfect.
• 8:08
  SPEAKER_01: Even if you lay out like That skill is two pages long, right?
• 8:14
  SPEAKER_01: Like there's a lot to it.
• 8:16
  SPEAKER_01: And I thought I had safeguards in place and it still gave me those questions.
• 8:22
  SPEAKER_01: So here's how I would tackle it next time.
• 8:26
  SPEAKER_01: Here's how I would try to amend that skill and get to explain what a skill is to change that.
• 8:32
  SPEAKER_01: But being clear with them about not just the successes,
• 8:39
  SPEAKER_01: but really The failures because of how rapidly it's changing and because
• 8:46
  SPEAKER_01: of what you can do with it has become really, really important.
• 8:51
  SPEAKER_01: So it's not just like the success stories.
• 8:54
  SPEAKER_01: It's here's where it failed me and I didn't catch it.
• 9:00
  SPEAKER_00: Yeah.
• 9:02
  SPEAKER_01: Bad on me.
• 9:03
  SPEAKER_01: You all get that point.
• 9:05
  SPEAKER_01: Congratulations, right?
• 9:07
  SPEAKER_02: Which is a practice that having taught similar classes or actually the same exact class.
• 9:14
  SPEAKER_02: Same exact class.
• 9:15
  SPEAKER_02: But even when I wrote them,
• 9:17
  SPEAKER_02: once in a while I would catch myself and I would go, oh, That was a bad item design.
• 9:22
  SPEAKER_02: I am sorry, everybody.
• 9:24
  SPEAKER_02: Here's where I went wrong.
• 9:25
  SPEAKER_02: And here's the point.
• 9:27
  SPEAKER_02: Yes.
• 9:27
  SPEAKER_02: And so absolutely.
• 9:29
  SPEAKER_02: That's, I think,
• 9:31
  SPEAKER_02: one of the magical things about teaching in education is there's always value
• 9:37
  SPEAKER_02: in opening up the process and making it visible.
• 9:41
  SPEAKER_02: It's sometimes hard.
• 9:44
  SPEAKER_02: It takes a little bit of honesty and not just honesty,
• 9:47
  SPEAKER_02: but A certain level of feeling secure enough to do that.
• 9:52
  SPEAKER_02: I don't know if it's easy for a TA in their first time running a class to say,
• 9:58
  SPEAKER_02: you know, I made a mistake last time.
• 10:01
  SPEAKER_02: Yeah, that is difficult.
• 10:03
  SPEAKER_02: But I think that after a while it becomes easier.
• 10:05
  SPEAKER_01: It is possible.
• 10:07
  SPEAKER_01: I mean, it is possible.
• 10:09
  SPEAKER_01: Because when we were working with particularly L.Ed.
• 10:14
  SPEAKER_01: students who envision these like perfect classrooms.
• 10:18
  SPEAKER_01: Yes.
• 10:19
  SPEAKER_01: I always want to be like, look, I was once one of you and I never had a perfect day.
• 10:24
  SPEAKER_02: Yeah.
• 10:25
  SPEAKER_01: Right.
• 10:26
  SPEAKER_01: And like that never happened.
• 10:28
  SPEAKER_01: But my kids all could read by the time they left my classroom.
• 10:33
  SPEAKER_01: My kids were all on benchmark when it came to math.
• 10:36
  SPEAKER_01: Like I was an effective teacher, but I never had a perfect day.
• 10:39
  SPEAKER_01: I messed up at least eight times every single day.
• 10:44
  SPEAKER_01: And you're gonna have to get used to that, right?
• 10:49
  SPEAKER_01: And I think that's a perfectionism thing that would be a separate,
• 10:53
  SPEAKER_01: separate... Definitely separate.
• 10:55
  SPEAKER_01: Separate issue.
• 10:56
  SPEAKER_02: Not that kind of daughter.
• 10:57
  SPEAKER_02: I seem to remember.
• 10:58
  SPEAKER_02: Yes.
• 10:58
  SPEAKER_01: Yes, because that's something... We did an episode of that.
• 11:01
  SPEAKER_01: That's something I think about a lot as well.
• 11:03
  SPEAKER_01: But so I think that the second part to your question was what I never want them to see me do.
• 11:09
  SPEAKER_00: Yeah, absolutely.
• 11:11
  SPEAKER_01: And I think that is...
• 11:13
  SPEAKER_01: I never want them
• 11:18
  SPEAKER_01: see me handing over my thinking to AI.
• 11:22
  SPEAKER_01: I never want them seeing that I am okay with just handing
• 11:31
  SPEAKER_01: over my pedagogical thinking for their learning over to an AI.
• 11:37
  SPEAKER_01: That, to me, is it.
• 11:40
  SPEAKER_01: That is it.
• 11:42
  SPEAKER_01: I am fine with them seeing me fail at using it.
• 11:46
  SPEAKER_01: I am fine at them seeing me struggle with it.
• 11:52
  SPEAKER_01: And that's also happened.
• 11:55
  SPEAKER_01: But within the successes,
• 11:59
  SPEAKER_01: within the failures,
• 12:00
  SPEAKER_01: within the parts where it's halfway there and it's not quite what I wanted it to be,
• 12:04
  SPEAKER_01: but that's what we're going to do today because class starts at 930, right?
• 12:09
  SPEAKER_01: Being able to talk through, here are the decisions that I've made.
• 12:14
  SPEAKER_01: Here's what I'm doing to ensure that you learn these concepts and here's why it's important.
• 12:24
  SPEAKER_01: I never want to abdicate that.
• 12:28
  SPEAKER_01: I don't want to abdicate my role as a teacher.
• 12:33
  SPEAKER_00: Yeah.
• 12:37
  SPEAKER_01: That's what I don't want them to see.
• 12:39
  SPEAKER_01: I don't want them to, if I ever say, oh, well, that's what AI did.
• 12:47
  SPEAKER_01: You know, that question, that question was what AI did.
• 12:51
  SPEAKER_01: So, whoops.
• 12:53
  SPEAKER_01: Like, my responsibility is to those students and to that learning.
• 13:01
  SPEAKER_01: And that's just me passing the buck.
• 13:04
  SPEAKER_01: So owning it, what I will do different next time, and making sure everyone gets credit for it.
• 13:09
  SPEAKER_01: That's what I never want them to see me do.
• 13:15
  SPEAKER_02: The other thing that is always on my mind,
• 13:17
  SPEAKER_02: because we talk a lot to students with the confidence of research,
• 13:22
  SPEAKER_02: so they tell us a few more things,
• 13:26
  SPEAKER_02: they talk about instructors reacting really badly to AI.
• 13:32
  SPEAKER_02: Yeah.
• 13:33
  SPEAKER_00: Mm-hmm.
• 13:33
  SPEAKER_02: Thinking, saying,
• 13:35
  SPEAKER_02: thinking,
• 13:36
  SPEAKER_02: behaving in very,
• 13:38
  SPEAKER_02: very charged ways that signal very strongly, not necessarily don't use AI.
• 13:45
  SPEAKER_02: What the instructors don't understand sometimes in the heat of the moment
• 13:49
  SPEAKER_02: that what it's really telling students is don't talk about AI.
• 13:53
  SPEAKER_02: Do what you do, don't talk about AI and don't get caught.
• 13:59
  SPEAKER_01: Which, what kind of world is don't get caught a good pedagogical plan?
• 14:06
  SPEAKER_02: And I don't think it comes out of a plan.
• 14:08
  SPEAKER_02: I think it happened especially in the first few semesters
• 14:11
  SPEAKER_02: when instructors got
• 14:12
  SPEAKER_01: their first clearly AI straight
• 14:16
  SPEAKER_02: out of ChatGPT output
• 14:18
  SPEAKER_02: that somebody didn't bother to do anything with.
• 14:21
  SPEAKER_02: And then you have this no AI, no way response.
• 14:28
  SPEAKER_02: Which I can understand where it came from, but it's not helpful.
• 14:31
  SPEAKER_01: Describe that to me because I can't understand it.
• 14:35
  SPEAKER_01: In what world is that an appropriate pedagogical response?
• 14:39
  SPEAKER_01: Why are you centering your discomfort with AI when it's not about you?
• 14:44
  SPEAKER_01: It's not about your learning.
• 14:46
  SPEAKER_01: It's not about your comfort.
• 14:47
  SPEAKER_02: Yeah.
• 14:48
  SPEAKER_01: It's about your learners.
• 14:49
  SPEAKER_01: And if that's where they're at, figure it out.
• 14:53
  SPEAKER_02: Yes.
• 14:54
  SPEAKER_02: And talk to them in a way because you don't want to close
• 14:57
  SPEAKER_01: that door.
• 14:57
  SPEAKER_01: You don't want them to hide from you.
• 15:00
  SPEAKER_01: That is unhelpful.
• 15:01
  SPEAKER_01: Right.
• 15:01
  SPEAKER_01: So I don't understand that response.
• 15:04
  SPEAKER_01: Yeah.
• 15:04
  SPEAKER_01: And that's what we don't want them to do.
• 15:06
  SPEAKER_01: You know, so I think about my online children's literature class.
• 15:10
  SPEAKER_01: Yeah.
• 15:11
  SPEAKER_01: Two summers ago, AI had or ChatGPT had gotten its first like big foothold where...
• 15:17
  SPEAKER_01: Students were using it writ large.
• 15:21
  SPEAKER_01: And I remember the first set of assignments were just slop, just such slop.
• 15:34
  SPEAKER_01: And there's two ways to respond to that,
• 15:37
  SPEAKER_01: which one is, oh my gosh, you use ChatGPT, naughty, naughty, naughty.
• 15:46
  SPEAKER_01: What does that do for anyone?
• 15:49
  SPEAKER_01: Versus it took a really long time, but like, hey, this is ChatGPT.
• 15:55
  SPEAKER_01: You can't tell me otherwise.
• 15:56
  SPEAKER_01: I know it's ChatGPT.
• 15:58
  SPEAKER_01: I'm not even going to argue with you around it.
• 16:02
  SPEAKER_01: I'm just going to take it as fact.
• 16:04
  SPEAKER_01: And the next time you do this, here's what I want you to do differently.
• 16:09
  SPEAKER_01: Because right now, here's how this is reading.
• 16:11
  SPEAKER_01: This is boring.
• 16:13
  SPEAKER_01: Boring AF.
• 16:15
  SPEAKER_01: And as someone who has to read 30 of these, like, that's not okay.
• 16:23
  SPEAKER_01: First off, the details, there are no details about the book that you supposedly read.
• 16:29
  SPEAKER_01: So, like,
• 16:29
  SPEAKER_01: here are the things that I need you to do next time you use this and walked them
• 16:35
  SPEAKER_01: through.
• 16:37
  SPEAKER_01: Because they're going, I mean,
• 16:38
  SPEAKER_01: it's unfortunately going to happen,
• 16:41
  SPEAKER_01: particularly in summer classes where things are Hectic and compressed.
• 16:47
  SPEAKER_01: And if you're seeing AI as a way to check things off your to-do list versus as
• 16:56
  SPEAKER_01: a tool for learning,
• 17:00
  SPEAKER_01: how can I tip that balance so that you're actually having to interact with
• 17:07
  SPEAKER_01: the things that I'm asking you to interact?
• 17:09
  SPEAKER_01: So this summer, I made them use ChatGPT And they submitted their dialogue with the chatbot.
• 17:17
  SPEAKER_01: I didn't care about the final product,
• 17:19
  SPEAKER_01: but it was that back and forth of,
• 17:22
  SPEAKER_01: okay, talk about this book with this ChatGPT thing.
• 17:26
  SPEAKER_01: And it went much better.
• 17:28
  SPEAKER_01: But that first summer,
• 17:33
  SPEAKER_01: the level of embarrassment that students were like, oh my gosh, I'm so embarrassed.
• 17:36
  SPEAKER_01: No, don't be embarrassed.
• 17:37
  SPEAKER_01: Just do this next time.
• 17:40
  SPEAKER_01: Don't hand me this.
• 17:43
  SPEAKER_01: And I
• 17:48
  SPEAKER_02: think, I mean, you talked about the fact that summer is more compressed.
• 17:51
  SPEAKER_02: It's also true,
• 17:52
  SPEAKER_02: I think,
• 17:52
  SPEAKER_02: for programs for people who are working full time and are also students,
• 17:56
  SPEAKER_02: whether they're doing their first degree or advanced degrees or whatever it is.
• 18:01
  SPEAKER_02: It's also true that it's different in online classes because you described this
• 18:05
  SPEAKER_02: in our previous episode where in your classroom where kids are coming,
• 18:09
  SPEAKER_02: you're giving them a piece of paper and they're writing on the paper and
• 18:12
  SPEAKER_02: then they're transferring it and they're playing with it.
• 18:15
  SPEAKER_02: But that's much harder to do when you're teaching asynchronously online.
• 18:20
  SPEAKER_02: Because they're already sitting at the computer.
• 18:22
  SPEAKER_02: They're already working only digitally.
• 18:25
  SPEAKER_02: So the snap between that and just going to ChatGPT is or to whatever.
• 18:30
  SPEAKER_02: I don't want to.
• 18:33
  SPEAKER_02: To give privilege,
• 18:34
  SPEAKER_02: but most people think about ChatGPT,
• 18:39
  SPEAKER_02: especially at that moment, and just cut and paste.
• 18:41
  SPEAKER_02: You're just cutting and pasting like you do with everything else,
• 18:44
  SPEAKER_02: and it is less and less you,
• 18:46
  SPEAKER_02: and I think you also emotionally care less and less after you do a few
• 18:49
  SPEAKER_02: of those runs.
• 18:51
  SPEAKER_01: Oh, yeah,
• 18:51
  SPEAKER_01: but I also think that if you have a set of,
• 18:58
  SPEAKER_01: okay, here's how many use this tool, right?
• 19:02
  SPEAKER_01: I like to think of it as ChatGPT or Gemini or Claude,
• 19:07
  SPEAKER_01: whichever one you're using, DeepSeek, I don't know if that's...
• 19:10
  SPEAKER_01: It's still there.
• 19:11
  SPEAKER_02: It's still in
• 19:12
  SPEAKER_01: there.
• 19:13
  SPEAKER_01: So whichever one you're using,
• 19:15
  SPEAKER_01: you want to make sure that the tool actually fits the function, right?
• 19:20
  SPEAKER_01: Yeah.
• 19:21
  SPEAKER_01: You can't screw in a screw with a hammer, right?
• 19:26
  SPEAKER_01: So just being aware of what that tool actually does,
• 19:30
  SPEAKER_01: it's not great at producing detailed summaries.
• 19:34
  SPEAKER_01: It just doesn't work that way.
• 19:36
  SPEAKER_01: So that explicit instruction at least gives them something interesting to be thinking about and
• 19:48
  SPEAKER_01: The other thing that I like about the way that I've set up
• 19:51
  SPEAKER_01: the AI instructions is it's iterative.
• 19:54
  SPEAKER_01: So the first round they do one thing,
• 19:56
  SPEAKER_01: the next round they do another thing,
• 19:58
  SPEAKER_01: and it forces them to be thinking about The back and forth,
• 20:03
  SPEAKER_01: which is, I think, where that learning comes from.
• 20:05
  SPEAKER_02: Yeah.
• 20:06
  SPEAKER_02: And that's the thing they need to see because they see us also structure going back
• 20:11
  SPEAKER_02: to what we want to see.
• 20:14
  SPEAKER_02: We want them to see the scaffolding that walks them step by step into something
• 20:18
  SPEAKER_02: and not just say, use this or don't use this for that matter.
• 20:22
  SPEAKER_01: And I mean,
• 20:23
  SPEAKER_01: I think it's going back to that notion of showing them The process,
• 20:27
  SPEAKER_01: like literally this is my cloud chat.
• 20:29
  SPEAKER_01: It is 82 pages of PDF for this one two-page document.
• 20:37
  SPEAKER_01: Like look at all of this setup that had to go into this.
• 20:43
  SPEAKER_01: But the document is pretty solid.
• 20:49
  SPEAKER_01: So disrupting that notion of, oh, this is an easy way to get work done.
• 20:53
  SPEAKER_01: It's not.
• 20:53
  SPEAKER_01: In fact, sometimes it's even harder.
• 20:55
  SPEAKER_01: Sometimes it takes more time.
• 20:56
  SPEAKER_01: So you have to think about it systematically.
• 21:00
  SPEAKER_01: Where is it going to be helpful?
• 21:01
  SPEAKER_01: Where is it going to save me time?
• 21:02
  SPEAKER_01: And where is it not?
• 21:03
  SPEAKER_01: Because sometimes it's just not.
• 21:06
  SPEAKER_01: It's not.
• 21:08
  SPEAKER_01: It's not.
• 21:08
  SPEAKER_02: I can't tell you how many times I go and say, hey, make me this presentation.
• 21:13
  SPEAKER_02: And I'm like, no.
• 21:17
  SPEAKER_02: I need to do a lot more prep work before I'm willing to take anything it does.
• 21:22
  SPEAKER_01: Yep.
• 21:22
  SPEAKER_01: The prep work is important.
• 21:25
  SPEAKER_01: Yeah.
• 21:26
  SPEAKER_01: And I think we don't.
• 21:28
  SPEAKER_01: It should already know all that, right?
• 21:29
  SPEAKER_01: Yeah.
• 21:30
  SPEAKER_02: No.
• 21:31
  SPEAKER_01: Isn't it all knowing?
• 21:32
  SPEAKER_01: Apparently.
• 21:34
  SPEAKER_01: No.
• 21:34
  SPEAKER_02: So to piggyback on that question and extend that a little bit, our students don't just watch us.
• 21:41
  SPEAKER_02: They're also going into schools where AI policies range from,
• 21:47
  SPEAKER_02: we turned on CHANT GPT last week and everybody's got access to including kids,
• 21:52
  SPEAKER_02: to You know,
• 21:55
  SPEAKER_02: complete bans,
• 21:56
  SPEAKER_02: nobody can use it for anything,
• 21:57
  SPEAKER_02: anywhere,
• 21:58
  SPEAKER_02: not just students,
• 21:59
  SPEAKER_02: teachers, administrators, nobody can use it, and anything in between.
• 22:05
  SPEAKER_02: And so, how do we prepare them to make good decisions about AI in the classroom,
• 22:12
  SPEAKER_02: when the rules are not something we control?
• 22:15
  SPEAKER_02: I mean, the rules are for somebody else to make,
• 22:17
  SPEAKER_02: and they will get a job where they will get a job.
• 22:23
  SPEAKER_02: Often, very often,
• 22:25
  SPEAKER_02: their mentor teachers,
• 22:27
  SPEAKER_02: whether they're in practicum or student teaching or even when they're teaching in
• 22:32
  SPEAKER_02: the first few years and they've got a mentor, might be well behind them in AI.
• 22:38
  SPEAKER_02: So that creates this really,
• 22:40
  SPEAKER_02: on one hand,
• 22:41
  SPEAKER_02: you've got the policy and then you've got what teachers are doing in their classroom
• 22:45
  SPEAKER_02: and what teachers are thinking.
• 22:47
  SPEAKER_02: And how do we prepare them to deal with all of these tensions?
• 22:51
  SPEAKER_01: Well, I think that If we focus
• 22:58
  SPEAKER_01: school policy is not something that we can control, right?
• 23:01
  SPEAKER_01: And I am a big fan of trying to just continually tell myself, what are the things I can control?
• 23:10
  SPEAKER_02: Mm-hmm.
• 23:11
  SPEAKER_01: And what are the things that I can't, right?
• 23:13
  SPEAKER_01: Like I can't control school policy, whether it's a free for all or whether it's a lockdown.
• 23:22
  SPEAKER_01: And so for me, in my practice, classroom facing policy, variance doesn't really matter to me.
• 23:33
  SPEAKER_01: What I have
• 23:36
  SPEAKER_01: I don't think that there's currently
• 24:04
  SPEAKER_01: A case where if you're not using student facing AI,
• 24:10
  SPEAKER_01: that those students are at a disadvantage yet.
• 24:13
  SPEAKER_01: And that could change.
• 24:15
  SPEAKER_01: That could very much change.
• 24:18
  SPEAKER_01: But for now, meh, right?
• 24:22
  SPEAKER_01: So like, do I need them to be in front of an AI enabled reading program?
• 24:28
  SPEAKER_01: Actually, I'd prefer they were not.
• 24:31
  SPEAKER_01: AI sucks at letter sound relationships,
• 24:34
  SPEAKER_01: but I think really the focus there for me is I can control how they use AI to prep.
• 24:44
  SPEAKER_01: And again, I think as with any tool, it comes down to function.
• 24:54
  SPEAKER_01: What does this tool allow me to do that would elevate what I'm already doing?
• 24:59
  SPEAKER_01: And I can always tell when a business
• 25:04
  SPEAKER_01: Is adopting AI simply because it's the cool thing to do and we want to be hip for
• 25:09
  SPEAKER_01: the young people and we're going to AI this and AI that.
• 25:13
  SPEAKER_01: And it serves no function.
• 25:16
  SPEAKER_01: Yeah.
• 25:16
  SPEAKER_01: Right?
• 25:17
  SPEAKER_01: If we think about for whom,
• 25:19
  SPEAKER_01: and there might be some people,
• 25:21
  SPEAKER_01: but for whom,
• 25:23
  SPEAKER_01: not AI, but similar, do the self-checkout lanes at Target make their life better?
• 25:30
  SPEAKER_01: Right?
• 25:34
  SPEAKER_01: But obviously it does.
• 25:36
  SPEAKER_02: No, not for me.
• 25:37
  SPEAKER_02: Not for you.
• 25:38
  SPEAKER_02: But I've been to a number of stores where there's a longer line for
• 25:42
  SPEAKER_02: the self-checkout than for...
• 25:45
  SPEAKER_02: So I am wondering about that every
• 25:48
  SPEAKER_01: time I go.
• 25:49
  SPEAKER_01: I would like to say that that is an indicator of a much larger problem that AI exacerbates.
• 25:56
  SPEAKER_01: Yes.
• 25:57
  SPEAKER_01: Where...
• 25:57
  SPEAKER_01: Yes, I agree.
• 25:58
  SPEAKER_01: We...
• 26:00
  SPEAKER_01: Are losing our spaces for the ability to construct an intersubjective reality, right?
• 26:09
  SPEAKER_01: Like if I am just engaging with this machine,
• 26:12
  SPEAKER_01: I'm not learning anything about Asher's day
• 26:16
  SPEAKER_01: or how his programming criminal justice goes or none of that.
• 26:22
  SPEAKER_01: Like he's graduating this semester and I know that because I will pick a human over
• 26:27
  SPEAKER_01: a machine that Any time.
• 26:30
  SPEAKER_01: Yeah.
• 26:31
  SPEAKER_02: Yeah.
• 26:31
  SPEAKER_01: Right.
• 26:32
  SPEAKER_01: So what does it say about us that there's a longer line for the self-checkout
• 26:39
  SPEAKER_01: than there is
• 26:41
  SPEAKER_02: for a person?
• 26:43
  SPEAKER_02: And I'd say the same thing because we've often heard students saying,
• 26:47
  SPEAKER_02: I would rather ask AI than go to office hours.
• 26:51
  SPEAKER_02: And I'm like, Office hours are important.
• 26:54
  SPEAKER_02: Connect with your instructor.
• 26:57
  SPEAKER_02: Talk to your teacher.
• 26:59
  SPEAKER_02: I know AI can do some of that work,
• 27:01
  SPEAKER_02: but there's so much else that goes on when you interact with a human being.
• 27:06
  SPEAKER_02: So that is the other side.
• 27:07
  SPEAKER_02: It's like, it can make me more efficient, but not at the expense of human connection.
• 27:13
  SPEAKER_01: Well, and when we're talking about AI, disrupting that myth that it's the expert.
• 27:20
  SPEAKER_01: Yeah.
• 27:21
  SPEAKER_01: Right?
• 27:22
  SPEAKER_01: Substack is a commonly held assumption that, oh, it just knows everything.
• 27:27
  SPEAKER_01: That's not how it works.
• 27:29
  SPEAKER_01: And so from a Vagatskin perspective, I'm the expert.
• 27:35
  SPEAKER_01: Come talk to me.
• 27:37
  SPEAKER_01: That's going to get you the most bang for your buck because you're paying a lot
• 27:43
  SPEAKER_01: of money for this class.
• 27:45
  SPEAKER_01: So you might as well use me while I'm here.
• 27:47
  SPEAKER_01: And Chat Chippy Tea is only $20 a month, if that.
• 27:52
  SPEAKER_01: Yeah, I don't even know.
• 27:53
  SPEAKER_01: Yeah, I'm still on the Claude Max plan.
• 27:56
  SPEAKER_01: Let's not talk about the shame.
• 27:57
  SPEAKER_01: Look, I'm single.
• 27:59
  SPEAKER_01: Okay.
• 27:59
  SPEAKER_01: I don't have kids.
• 28:00
  SPEAKER_01: I have three dogs.
• 28:01
  SPEAKER_01: What else do I have in my life?
• 28:02
  SPEAKER_01: Don't judge.
• 28:04
  SPEAKER_01: Don't judge.
• 28:05
  SPEAKER_01: But like
• 28:09
  SPEAKER_01: when they're thinking about function and everything
• 28:15
  SPEAKER_01: in our school day should be about function.
• 28:20
  SPEAKER_01: How is that helping push learning?
• 28:22
  SPEAKER_01: How is that giving them opportunities to engage in retrieval practice?
• 28:28
  SPEAKER_01: How is that giving them an opportunity to add nuance to an existing conceptual map,
• 28:33
  SPEAKER_01: et cetera, et cetera?
• 28:36
  SPEAKER_01: AI can help with those things, but you have to be really specific.
• 28:43
  SPEAKER_01: In terms of how it's helping you.
• 28:45
  SPEAKER_01: So it's a really interesting testing ground for how are you understanding the pedagogy, right?
• 28:53
  SPEAKER_01: So to go back to that modeling piece, like modeling is tricky, right?
• 28:58
  SPEAKER_01: Like if we start with the scenario of you're teaching someone to cross the street,
• 29:04
  SPEAKER_01: what do you do?
• 29:07
  SPEAKER_01: And it's typically I look at the walk sign, right?
• 29:12
  SPEAKER_01: And once I see the walking person, I walk across.
• 29:18
  SPEAKER_01: And it's like, ooh, do you really want to do that?
• 29:20
  SPEAKER_01: Because that light is just the legal indicator that you could cross right now.
• 29:27
  SPEAKER_01: You probably still want to look to the left and the right and then look to
• 29:30
  SPEAKER_01: the left again really quickly because your gaze was diverted.
• 29:34
  SPEAKER_01: And then cross.
• 29:35
  SPEAKER_01: And where do you cross?
• 29:37
  SPEAKER_01: And what kind of clip?
• 29:38
  SPEAKER_01: Et cetera, et cetera.
• 29:40
  SPEAKER_01: And When they go back into ChatGPT or Claude or Gemini and they ask about like,
• 29:50
  SPEAKER_01: this is someone who doesn't know how to do this.
• 29:52
  SPEAKER_01: Here's the task.
• 29:54
  SPEAKER_01: What am I missing?
• 29:56
  SPEAKER_01: And it's really helpful with things like that, but it's the function of I need better language.
• 30:04
  SPEAKER_01: I am an adult who does this automatically.
• 30:07
  SPEAKER_01: Help me defossilize this.
• 30:09
  SPEAKER_01: How do I do that?
• 30:12
  SPEAKER_01: And additionally, to go back to the mentor teachers, mentor teachers can be wrong.
• 30:23
  SPEAKER_01: Just as our students can be wrong.
• 30:24
  SPEAKER_01: We can be wrong.
• 30:26
  SPEAKER_01: And part of,
• 30:28
  SPEAKER_01: I think,
• 30:28
  SPEAKER_01: what we need to be doing is helping them build enough professional confidence,
• 30:36
  SPEAKER_01: not arrogance,
• 30:37
  SPEAKER_01: just confidence,
• 30:39
  SPEAKER_01: to know When to adopt,
• 30:42
  SPEAKER_01: when to hold a line, when to be like, I don't need AI for this.
• 30:49
  SPEAKER_01: I can do it in different ways.
• 30:53
  SPEAKER_01: And I think that's where the transmission,
• 30:56
  SPEAKER_01: that's where the transfer comes from, regardless of your policy at your school.
• 31:03
  SPEAKER_02: All right.
• 31:05
  SPEAKER_01: For agreeing to do a second episode.
• 31:07
  SPEAKER_01: Exactly.
• 31:08
  SPEAKER_00: Thank you so much.
• 31:09
  SPEAKER_01: Very welcome.
• 31:10
  SPEAKER_01: I didn't.
• 31:11
  SPEAKER_01: Yeah.
• 31:11
  SPEAKER_01: Thanks.
• 31:12
  SPEAKER_01: This was fun.
• 31:13
  SPEAKER_01: All right.
• 31:13
  SPEAKER_02: Hopefully I didn't sound like an idiot.
• 31:16
  SPEAKER_02: That's my goal.
• 31:18
  SPEAKER_02: For the second time.
• 31:19
  SPEAKER_02: We set the bar.
• 31:20
  SPEAKER_01: I mean, is that low?
• 31:23
  SPEAKER_01: Is that high?
• 31:24
  SPEAKER_01: I don't know.
• 31:24
  SPEAKER_01: I mean, look.
• 31:25
  SPEAKER_01: Every day.
• 31:25
  SPEAKER_01: First two things on my to-do list.
• 31:27
  SPEAKER_01: Don't kill the dogs.
• 31:29
  SPEAKER_01: And don't be an idiot.
• 31:31
  SPEAKER_02: Check.
• 31:32
  SPEAKER_01: At least one of them.
• 31:33
  SPEAKER_02: I don't know how the dogs are doing.
• 31:34
  SPEAKER_02: I'll see when I get home.
• 31:36
  SPEAKER_02: All right.
• 31:36
  SPEAKER_02: Thank you.
• 31:37
  SPEAKER_02: And we'll see you next time on... Azadeh.
• 31:40
  SPEAKER_00: And Guy.
• 31:40
  SPEAKER_00: On AI.
• 31:42
  SPEAKER_02: And this time it went wrong.