Not That Kind of Doctor - Online Teaching

TLTE Author

04/04/2025 Added

0 Plays

Description

Online Teaching - Not That Kind of Doctor with Nick Husbye and Guy Trainin www.youtube.com/@tltenotthatkindofdoctor

Searchable Transcript

Search:

[00:00:00.750]So, you thought online teaching
[00:00:03.660]would give you flexibility.
[00:00:05.700]Instead, your MLS keeps emailing students at 2:00 AM
[00:00:09.450]and your grading discussion posts
[00:00:11.070]like a customer service bot.
[00:00:12.930]And now the bots want to teach, too.
[00:00:16.020]Welcome to the Post Human Classroom.
[00:00:18.810]Today we are talking about the fine line between help
[00:00:22.200]and harm when it comes to AI in online teaching.
[00:00:26.940]What to use, what to skip,
[00:00:28.920]and how to avoid becoming a digital ghost.
[00:00:31.710]And if you're wondering how this will affect
[00:00:33.750]your tenure file, don't worry, we've got you covered.
[00:00:39.990]I'm Nick Husbye, Associate Professor
[00:00:41.670]of Elementary Literacy Education here at UNL.
[00:00:45.030]And I'm Guy Trainin, Professor of Education here at UNL.
[00:00:48.480]And this is "Not That Kind Of Doctor."
[00:00:51.500](upbeat electronic music)
[00:00:59.910]How you doing? I'm doing okay.
[00:01:01.895]Yeah? Yeah.
[00:01:03.000]Okay, that's good. That's good.
[00:01:04.968]It's late in the semester.
[00:01:07.320]Upright, not crying. Um, no.
[00:01:10.140]Oh.
[00:01:11.396]But really not thinking about how many assignments
[00:01:13.020]I have to give feedback on, so yes.
[00:01:15.480]Okay, so, you know, delusion.
[00:01:18.484]Yes, delusion.
[00:01:19.317]Delusion is the way forward.
[00:01:20.150]Delusion is where I'm at.
[00:01:20.983]Delusion is the way forward.
[00:01:21.816]Okay. Yes.
[00:01:22.649]I appreciate that.
[00:01:23.482]I'm in a phase where I am rethinking
[00:01:27.150]my online class for this summer.
[00:01:28.830]Mm-hmm.
[00:01:29.663]And so one thing that I've been trying
[00:01:34.470]to do is think through how to thoughtfully leverage
[00:01:39.090]AI resources in helping me develop
[00:01:44.340]a more robust, a more engaging-
[00:01:47.010]Mm-hmm.
[00:01:47.935]A more thoughtfully designed online course.
[00:01:51.900]'Cause that's when I teach this course is in the summer.
[00:01:54.190]Yeah.
[00:01:55.023]And it's asynchronous and online.
[00:01:56.910]And so I've been thinking about some of the ways
[00:02:01.440]that I can leverage AI to create presence, right?
[00:02:06.994]Yes.
[00:02:07.864]Because when you're thinking about online courses,
[00:02:09.330]you're not showing up in person, right?
[00:02:11.070]You're in videos,
[00:02:14.130]you're in comments,
[00:02:15.420]you're sending out announcements,
[00:02:17.220]et cetera, et cetera, et cetera.
[00:02:20.550]So I would love,
[00:02:22.860]because I know AI just tickles you, it makes you so happy.
[00:02:25.230]It does make me happy.
[00:02:26.063]So maybe this is what you need this morning.
[00:02:27.060]And sometimes worried, so.
[00:02:28.474]Sometimes worried, yeah, okay.
[00:02:29.760]Yeah.
[00:02:30.600]Same coin, different size.
[00:02:31.980]Mm-hmm.
[00:02:32.813]So maybe this is what you need.
[00:02:34.890]Okay. To cheer up a little bit,
[00:02:36.090]because I would love to know.
[00:02:37.964](laughing) Yes.
[00:02:39.030]As you're thinking about the potentials of AI
[00:02:42.120]and all the playing that you've done with AI.
[00:02:43.648]Mm-hmm?
[00:02:44.730]What's some ways that we could use AI
[00:02:49.050]to help students know that there's actually
[00:02:50.400]a real person behind their screen?
[00:02:52.440]What do you see as some of the affordances,
[00:02:54.840]as well as the limitations, between all of that?
[00:02:58.440]So, really, the thing that I think about AI
[00:03:05.430]is actually trying to generate a little bit of presence
[00:03:10.639]in the class by automating, not automating,
[00:03:15.030]but delivering some responses.
[00:03:17.070]I'm a little bit hesitant about saying automating,
[00:03:19.290]because automating always means
[00:03:21.510]if A, then B definitely follows.
[00:03:23.910]And one of the things that AI does
[00:03:25.590]is it makes some judgment calls about what can happen.
[00:03:29.310]So for example, one of the things that can be done
[00:03:33.120]is creating a chatbot that responds to questions,
[00:03:39.240]or practice for an exam, or a midterm, whatever assignment,
[00:03:45.660]that is based on a corpus
[00:03:47.070]that's directly connected to the class.
[00:03:48.870]And so there's been a very successful deployment
[00:03:51.510]of these bots that are trained,
[00:03:55.320]and have a very specific corpus that you can use.
[00:03:58.440]Whether it's something like NotebookLM, for example, does,
[00:04:01.140]you upload a few documents, and then you can chat
[00:04:03.810]with the documents, getting responses,
[00:04:07.320]or potentially even getting questions from the bot,
[00:04:10.260]and saying, "Here are 10 questions that might appear
[00:04:12.869]on a quiz or an exam or an essay," or whatever that is.
[00:04:20.700]And so those, for me, are really, really promising,
[00:04:23.760]because students can get real responses.
[00:04:27.720]Most of the data that we have right now
[00:04:29.430]about these practices is they seem to be very solid.
[00:04:32.640]So students sometimes try to push the bot, right?
[00:04:36.900]What can I get the bot to say?
[00:04:38.760]But it turns out that the bots seldom say
[00:04:41.640]something that is outside the basic data
[00:04:44.730]they're supposed to refer back to.
[00:04:46.800]And usually what they do, if they're well designed,
[00:04:49.470]is they start saying very general things,
[00:04:52.230]like when they say, you know,
[00:04:53.857]"Pretend that you are now something,
[00:04:56.490]and respond to this material."
[00:04:58.410]And so they give this very vague,
[00:05:00.180]very ChatGPT-ish answer, in some ways.
[00:05:02.536]Okay.
[00:05:03.369]That it's like, yeah, that's not helpful.
[00:05:05.670]But it's also not crossing boundaries.
[00:05:07.380]They make very, very few mistakes.
[00:05:09.090]And part of the conversation with students is,
[00:05:11.310]of course, this is a chatbot.
[00:05:13.890]It's based on the materials. Mm-hmm.
[00:05:16.020]Always check your materials.
[00:05:17.970]Well, and so I'm curious about, like,
[00:05:20.340]how hard is it to build a chatbot?
[00:05:21.810]'Cause I actually haven't done that yet.
[00:05:24.150]It can be fairly simple.
[00:05:26.160]And, um-
[00:05:27.960]Like how simple are we talking?
[00:05:29.130]Like, knitting a washcloth versus knitting a sweater?
[00:05:35.280]Knitting a wash cloth?
[00:05:36.510]I think that it depends on what you have available
[00:05:38.790]at your institution.
[00:05:39.660]So the first thing- Okay.
[00:05:40.620]You need to say is, "Dear institution,
[00:05:43.530]what do you have available?"
[00:05:44.880]Dearest institution. Yes.
[00:05:46.740]Dost thou have thy resources?
[00:05:49.140]Yes.
[00:05:50.490]Please respond in the language of Shakespeare.
[00:05:54.720]So those are, many of those are fairly easy to make,
[00:06:00.120]and you do not need, so maybe what I'm reading
[00:06:03.150]into your question is do I need to know how to program?
[00:06:05.640]And the answer is no. Perfect.
[00:06:08.190]This is really important. We love that.
[00:06:09.780]There are sets, and some of the LMSs are trying
[00:06:14.070]to start giving the option to do that.
[00:06:18.960]And what you do is you put the materials,
[00:06:22.237]and upload the materials,
[00:06:24.540]and they can be a recording of the videos you've made.
[00:06:28.170]Mm-hmm.
[00:06:29.003]They can be the textbook,
[00:06:32.340]or open materials that you're using,
[00:06:36.600]and then it processes them,
[00:06:40.020]and uses the bigger, large language model
[00:06:45.180]to produce actually text that is meaningful.
[00:06:48.030]Okay.
[00:06:48.999]So the large language models really don't contribute
[00:06:50.700]a lot of the knowledge, but they contribute the syntax,
[00:06:53.580]the ability to chat in normal language,
[00:06:56.160]and not to just give quotes from the books.
[00:06:59.190]And if they're well designed, they also provide links
[00:07:02.040]to the original material.
[00:07:03.360]So you can actually, and much like right now,
[00:07:05.970]if you use ChatGPT, deep knowledge,
[00:07:10.080]where you do the, deep research, sorry.
[00:07:12.349]After each piece, there's a link that says,
[00:07:15.967]"Here's where the information came from."
[00:07:17.820]So you can go and look for yourself
[00:07:19.350]and make sure that it's actually accurate.
[00:07:21.360]Okay.
[00:07:22.193]So it's a much more secure from a data perspective.
[00:07:29.100]So that's one.
[00:07:30.000]The second thing is if you are going to integrate it
[00:07:33.000]into your LMS, you must make sure that you know
[00:07:35.490]what your institution guidelines and policies are.
[00:07:38.155]Mm-hmm. About that.
[00:07:39.120]And I highly suggest that you also ask your students
[00:07:43.680]to do whatever training is available.
[00:07:45.630]So for example, at UNL, there's training available
[00:07:47.940]for every student and faculty member
[00:07:50.190]about using any large language model in AI.
[00:07:54.840]And it talks about privacy,
[00:07:56.700]it's a very short module.
[00:07:58.080]I think it takes half an hour to 40 minutes.
[00:08:00.360]But it's really important
[00:08:01.950]because then students understand what's in there,
[00:08:04.110]and also what data not to put into a chatbot.
[00:08:07.560]Right? Right.
[00:08:08.430]And it's also good for instructors to know
[00:08:11.250]what the boundaries are.
[00:08:12.390]So for example, if you are using something
[00:08:16.230]that is third party, and you don't know
[00:08:18.000]where the data is hosted,
[00:08:19.230]you really don't want to use it in any way
[00:08:22.170]to look at student products,
[00:08:24.240]because that data is now, you're taking what is
[00:08:28.321]inherently students' copyright, because they've created it.
[00:08:32.600]Right.
[00:08:33.433]Or at least we hope they did.
[00:08:35.550]And then you send it,
[00:08:36.840]and now somebody else has control over it,
[00:08:38.970]and might be using it to do training, and all of that,
[00:08:41.460]and it's not yours to share.
[00:08:42.960]Right.
[00:08:43.793]So, that's one component.
[00:08:46.050]And you can think about the other components.
[00:08:48.357]The most useful place is really in creating those bots
[00:08:52.200]that respond based on course material.
[00:08:54.900]I think that creates that.
[00:08:56.940]If I want to ask questions before a midterm, a bot is great,
[00:09:01.260]because a bot doesn't get tired, it will answer at midnight
[00:09:04.830]where we talk about boundaries.
[00:09:07.290]If you're teaching online, and you want presence-
[00:09:09.301]Right.
[00:09:10.560]You do not want to create the situation
[00:09:13.230]where you're responding at all times.
[00:09:14.760]So those are the things that I think about,
[00:09:16.770]as you think about.
[00:09:17.970]Yeah. What to do.
[00:09:19.028]You don't wanna be available 24/7.
[00:09:21.030]Yes. Right?
[00:09:21.863]Because that's not any way to live.
[00:09:24.150]Mm-mm.
[00:09:24.983]It's just-
[00:09:25.816]But the temptation, I mean, the flip side of this
[00:09:28.530]is the temptation is, this is how you create presence,
[00:09:31.590]by being hyper-responsive.
[00:09:33.210]If I respond, and I have some colleagues who respond
[00:09:36.090]to students within 20 minutes.
[00:09:38.790]Sometimes within five minutes.
[00:09:40.260]That's not me.
[00:09:41.187]And that's not me, either. (Nick chuckles)
[00:09:43.800]Although I try to be really highly responsive.
[00:09:46.230]But I think it creates unreasonable expectations,
[00:09:48.630]because once that you are known for that,
[00:09:52.410]students are starting to read the fact that you've,
[00:09:55.140]you know, you've just taken an hour to make dinner
[00:09:57.300]and spend time with family, or with friends,
[00:09:59.730]or whatever, as something is wrong with what I just did,
[00:10:03.450]or they don't like me.
[00:10:04.890]I mean, you are creating a whole host of outcomes
[00:10:07.980]that are unhealthy, for you or for others,
[00:10:11.130]when you do that.
[00:10:12.000]So the bots can be a replacement that isn't you.
[00:10:16.440]And I've seen some people playing, actually,
[00:10:18.534]with a live avatar of you.
[00:10:22.925]Oh.
[00:10:23.762]So you create an avatar of yourself.
[00:10:25.110]I just came back from a conference,
[00:10:26.550]and one of the researchers there did that
[00:10:30.807]and there was a long discussion about the ick factor.
[00:10:34.560]It was disconcerting to everybody there.
[00:10:38.490]Yeah.
[00:10:39.323]To everybody trying to consume this,
[00:10:40.770]here's your professor, but it's not your professor,
[00:10:42.990]and it's his face and his voice,
[00:10:44.370]but he's talking somewhat incoherently,
[00:10:46.920]and it doesn't sound quite like him.
[00:10:49.260]So it's uncanny valley kind of an effect.
[00:10:52.590]Unhealthy.
[00:10:53.423]Don't do that.
[00:10:54.690]Yeah, no. At least not yet.
[00:10:55.860]No.
[00:10:56.693]Don't do that. Don't do that.
[00:10:58.320]Because that goes back, that goes into our second point.
[00:11:02.880]And that is?
[00:11:05.190]It's not a replacement. It's not a replacement.
[00:11:07.440]AI is not a replacement.
[00:11:09.390]And what we're seeing in the research
[00:11:11.520]is AI is very good.
[00:11:12.960]If you train a bot like that,
[00:11:14.460]it is very good at staying within the material,
[00:11:18.510]which is what you want it to do.
[00:11:20.640]But giving examples from other times, and other places,
[00:11:24.180]or other topics is where the instructor is instrumental,
[00:11:28.920]because LLMs can make that jump,
[00:11:31.560]unhealthily, and often with many mistakes.
[00:11:33.750]This is exactly why we bound these bots to say no.
[00:11:36.810]You cannot go much beyond what we're giving you.
[00:11:40.798]And also, the bots are not very good at understanding
[00:11:45.480]what people misunderstand.
[00:11:49.320]So instructors, because we have taught this course
[00:11:52.410]many times, we have a notion about where do people
[00:11:55.530]have a hard time?
[00:11:57.060]And we work around that, and we have a set of examples,
[00:12:00.990]or ways to explain things.
[00:12:03.270]So for example, in a recent article
[00:12:07.140]that had a bot that was trained on statistics,
[00:12:10.230]it was very good at giving some examples,
[00:12:14.100]and explaining some phenomena,
[00:12:15.960]but it had really a hard time explaining other phenomena.
[00:12:21.180]So for example, it was working on correlation and causation,
[00:12:24.810]and it was very good in explaining why correlation,
[00:12:29.010]and to give an example, why correlation
[00:12:30.870]and causation are not the same.
[00:12:32.370]But trying to explain what might be a third factor
[00:12:36.060]that is causing both of them, that it did not.
[00:12:39.780]It was not able to produce, so.
[00:12:41.640]Interesting.
[00:12:42.570]So it has moments
[00:12:46.710]where it just doesn't have the repertoire
[00:12:49.350]of somebody who's been teaching a long time, right?
[00:12:52.017]Mm-hmm.
[00:12:52.850]And has learned that this specific examples
[00:12:56.400]really help clarify things for everybody.
[00:12:59.250]And if you take then some of your materials
[00:13:02.610]and feed it back to the bot, maybe it'll get better.
[00:13:05.490]But also there's always that ability to read
[00:13:10.620]your students in a way,
[00:13:13.080]and understand where they're coming from.
[00:13:14.790]Which is, again, for bots, is considerably harder.
[00:13:20.160]Yeah, definitely.
[00:13:21.630]But I do like it to help me
[00:13:23.970]with, like not children's literature.
[00:13:26.490]Yeah.
[00:13:27.323]But I'm working on revising
[00:13:30.180]the phonics fund foundations of reading stuff.
[00:13:32.640]Yeah.
[00:13:33.593]For literacy block right now.
[00:13:34.860]And so I know where students
[00:13:36.870]kind of run into these roadblocks.
[00:13:39.510]I have documented that.
[00:13:41.137]Mm-hmm.
[00:13:41.970]And what I'm finding AI has really been helpful
[00:13:45.540]with is, oh hey,
[00:13:48.810]here's some different ways to explain it.
[00:13:50.897]Yes.
[00:13:51.769]Here's where you're lacking teacher clarity.
[00:13:53.280]And that's been interesting.
[00:13:55.200]Yes.
[00:13:56.033]So like, revising scripts,
[00:13:58.650]or thinking about different ways to,
[00:14:03.030]this is how I scaffolded it previously,
[00:14:05.100]this is how I'm gonna scaffold it this time.
[00:14:07.230]Mm-hmm.
[00:14:08.063]That's been really helpful.
[00:14:09.810]But then the examples suck.
[00:14:13.710]Yes.
[00:14:15.082]And one of the great things about AI,
[00:14:17.730]and this is my favorite thing maybe,
[00:14:19.560]is that AI will generate 50 ideas,
[00:14:22.740]or 20 ideas, 50 is probably too much.
[00:14:24.930]But it can generate 20 ideas,
[00:14:26.790]and out of those 20, you choose the one you like.
[00:14:29.346]Mm-hmm.
[00:14:30.179]It is not excellent at choosing the best one,
[00:14:32.700]but it is excellent at generating 20.
[00:14:35.010]It takes us a long time to generate 20
[00:14:38.370]completely separate ideas,
[00:14:39.600]because we kind of start thinking about
[00:14:41.820]our favorite idea, and start running with it.
[00:14:44.130]And this allows at the ideation phase
[00:14:47.370]really to open things up.
[00:14:49.331]I use AI right now to look at my syllabi,
[00:14:52.380]and identify where I don't have enough detail.
[00:14:55.110]Mm-hmm.
[00:14:56.070]Though I find that when you ask it to provide
[00:14:58.980]some of that detail, it often overdoes it
[00:15:01.860]in a way that I find unhelpful.
[00:15:04.290]So there's such a level of detail
[00:15:06.210]that it's actually creating those step by steps,
[00:15:10.380]for students that are so rigid
[00:15:13.200]that it's not really the kind of product I wanna see.
[00:15:17.040]Right.
[00:15:17.873]Well, and part of that is you have to tell them,
[00:15:20.130]Yeah.
[00:15:21.480]You have to tell the AI,
[00:15:23.557]"These are the characteristics of what I'm looking for.
[00:15:26.310]This is what I like about my current system."
[00:15:28.320]Mm-hmm.
[00:15:29.153]I mean, it's really like training.
[00:15:30.030]Yes. Right?
[00:15:33.077]'Cause one of the other things that I've been doing with AI,
[00:15:35.850]as I'm thinking about the flow of this asynchronous
[00:15:38.760]online class- Mm-hmm?
[00:15:41.100]Is here's all the materials I've created,
[00:15:44.790]here's kind of what the flow is.
[00:15:47.730]Talk to me about what bridges haven't I built
[00:15:51.360]as students move from activity to reading
[00:15:54.150]to assignment to reading to activity.
[00:15:57.330]And that's been really helpful.
[00:15:59.280]Yes.
[00:16:00.113]In ways that I
[00:16:04.350]have not always been as great at thinking through
[00:16:07.601]as I would like to be.
[00:16:09.960]And so it's not doing that work for me.
[00:16:14.190]I'm still, like, it's making me-
[00:16:17.800]That's true. More efficient,
[00:16:18.960]but maybe more efficient by like,
[00:16:23.070]maybe it takes me 75% of the time,
[00:16:26.760]whereas it took me 100% of the time earlier.
[00:16:29.790]So it's not like a huge cost or time savings,
[00:16:33.690]but it's still a time savings.
[00:16:35.610]And what I find that those time savings
[00:16:39.240]are actually motivating me to act sooner.
[00:16:42.030]Mm-hmm.
[00:16:42.960]That is, I know there's help.
[00:16:45.120]So the initial step is done with AI,
[00:16:47.940]and then I'm already in the task,
[00:16:49.470]and it's gonna get done, and it's gonna get done quickly.
[00:16:52.650]So yes, the midcourse, or the just in time adjustments.
[00:16:59.790]So you can do some of this in advance,
[00:17:01.470]but you can do also just in time adjustments using AI
[00:17:05.070]while you're teaching, being more sensitive,
[00:17:08.310]and more readily available for your students,
[00:17:11.670]because you're seeing certain effects.
[00:17:13.680]Oh, I've just identified that this group,
[00:17:16.740]I teach the introduction to educational inquiry
[00:17:21.990]once in a while online.
[00:17:23.730]I don't love it, but- (Nick laughing)
[00:17:26.760]I don't love doing things online.
[00:17:28.200]I'll be honest.
[00:17:29.033]I'm all digital.
[00:17:30.390]That part is less fun for me. Interesting.
[00:17:33.120]Much less fun for me. Huh.
[00:17:34.770]I love the interaction,
[00:17:35.910]and online, you don't have interaction and I-
[00:17:39.196]You could, you just haven't thought through.
[00:17:40.950]Well, no.
[00:17:41.783]And I have implemented interaction throughout
[00:17:45.210]in different ways,
[00:17:46.710]but the way that I do is trying to realize
[00:17:52.500]who my students are in an introductory course like that,
[00:17:55.590]it's really very different from semester to semester.
[00:17:58.920]So some semester, the majority
[00:18:00.660]of our students are getting the master's
[00:18:02.520]with a teaching degree, with a teaching certificate.
[00:18:05.280]So they are very much at the beginning of their journey.
[00:18:08.580]Whereas sometimes we have a lot of veteran teachers
[00:18:13.170]who are really,
[00:18:16.590]have years of experience,
[00:18:18.240]so their thinking is considerably different.
[00:18:21.030]So adjusting based on that.
[00:18:22.440]One time I had two priests in my class.
[00:18:24.728]Oh, fun.
[00:18:25.710]That were also educators.
[00:18:27.570]They were lovely.
[00:18:28.620]That was- I'm sure.
[00:18:31.227]A fantastic experience,
[00:18:33.900]as their wisdom came in different ways
[00:18:36.570]to inform other people.
[00:18:37.590]It was lovely.
[00:18:38.520]And so adjusting to that in real time,
[00:18:42.210]AI can help you think through
[00:18:44.040]what might be a way to address this issue?
[00:18:46.950]And again, give me options.
[00:18:48.780]Okay, I'm gonna go with this,
[00:18:50.070]or I don't like any of these, but now I'm thinking.
[00:18:53.070]Right.
[00:18:53.903]And we love the thinking. Yes.
[00:18:54.736]We love the thinking partner.
[00:18:55.569]That's probably my favorite part
[00:18:56.850]of any AI chatbot right now is-
[00:19:00.666]Mm-hmm, yes.
[00:19:02.490]I've got mine trained to be like,
[00:19:05.077]"Nick, that's a stupid idea."
[00:19:07.409](Guy laughing)
[00:19:10.035]That's the kind of person I am.
[00:19:11.088]I would like my AI to be nicer to me.
[00:19:13.530]No, I don't know.
[00:19:15.180]Because then they're like, "Well, that's an okay idea."
[00:19:18.060]I don't want that.
[00:19:19.020]I don't have time.
[00:19:20.130]I'm trying to get more efficient.
[00:19:21.510]Tell me when it sucks, and let's move on.
[00:19:25.050]I don't need my ego stroked.
[00:19:27.000]That is true. By AI.
[00:19:28.290]That is true.
[00:19:29.640]That is true.
[00:19:30.665]I actually don't, either, but it sometimes-
[00:19:33.020]It seems you like it.
[00:19:34.290]You like it when your AI is nice to you.
[00:19:37.844]Nah, no I-
[00:19:38.677]Is it because no one else in your life is nice to you?
[00:19:40.920]No, people are nice to me.
[00:19:42.210]Should I be worried? I'm good.
[00:19:43.598]No. Okay.
[00:19:44.431]I am just fine. Just checking.
[00:19:45.750]Alright, let's talk about some red flags.
[00:19:48.990]Oh, some red flags. Yes.
[00:19:50.700]Well, we've talked about some red flags already.
[00:19:52.200]Yeah.
[00:19:53.432]There's the ethical considerations
[00:19:54.265]of, like, what you're feeding
[00:19:55.740]into ChatGPT. Yes.
[00:19:57.150]So like, if it's stuff that,
[00:20:04.080]you know, that is copyrighted by someone.
[00:20:06.060]Yeah, yeah.
[00:20:07.950]Might not want to feed that in there.
[00:20:09.600]Yeah.
[00:20:10.433]I am finding it interesting that sometimes
[00:20:11.910]I'll be like, "What do you know about this article?"
[00:20:14.490]Yeah.
[00:20:15.323]And I'll just get the title of the article,
[00:20:17.190]and it just spits out the article.
[00:20:19.560]Yeah. That's interesting to me.
[00:20:21.390]I did not input that article.
[00:20:22.800]Yeah.
[00:20:24.300]But that article was scraped at one point or another.
[00:20:27.187]It was.
[00:20:28.058]It was, and then it puts you in this weird, like,
[00:20:29.790]well, do I leverage that or do I not?
[00:20:32.250]Hmm.
[00:20:34.170]Well, if you keep it within the LMS.
[00:20:37.068]Eek.
[00:20:38.160]Well, I don't wanna use the article.
[00:20:39.540]I want to use it to inform- Yeah.
[00:20:41.580]Things around the article.
[00:20:43.290]Right? Yeah.
[00:20:44.123]So that is an ethical weird space for me,
[00:20:47.986]because it's already part of the knowledge base,
[00:20:51.480]does that mean I should use it,
[00:20:53.070]is kind of my larger question.
[00:20:56.490]I also, we've talked about auto replying.
[00:20:59.280]Yeah.
[00:21:00.113]Do not.
[00:21:01.200]Content creation is interesting,
[00:21:06.720]particularly from when you're using ChatGPT
[00:21:09.960]for content creation.
[00:21:11.700]As long as you have some, it's like bowling,
[00:21:14.430]and you have the bumpers, right?
[00:21:15.720]Yeah.
[00:21:16.553]As long as you have some bumpers on it,
[00:21:17.386]like, this is what I'm doing, this is not what I'm doing.
[00:21:19.350]This is the tone of where we are,
[00:21:20.790]this is not the tone of where we are.
[00:21:23.653]'Cause it can get super complex.
[00:21:26.550]Like, "Here's a script for your video.
[00:21:28.950]It's a cinematic overview (Guy laughs)
[00:21:30.780]when you're hovering over the city,"
[00:21:32.520]and I'm like, I barely know
[00:21:35.820]how to turn on Premier Pro, much less-
[00:21:40.650]Yes. Do all of this stuff.
[00:21:42.360]And so that, for me, is part of it.
[00:21:45.810]But also, from a content creation perspective,
[00:21:48.480]there's a lot of slop.
[00:21:50.430]Oh yeah.
[00:21:51.480]And that is, there are all of these engines to create.
[00:21:54.870]I will create your PowerPoint.
[00:21:56.040]Just tell us what it's about.
[00:21:57.240]It's never good. Oh, they're never good.
[00:21:58.540]It is never good. Never good.
[00:22:00.810]Maybe one day, it'll get good.
[00:22:02.460]Right now, it is not good.
[00:22:04.500]And I play with it once in a while,
[00:22:06.660]because I'm like, maybe it's improved.
[00:22:08.190]Last time you, were playing with a video.
[00:22:09.930]Yeah. An explainer video.
[00:22:11.538]And so can you create something?
[00:22:14.700]Absolutely.
[00:22:15.570]But auto-create is never good.
[00:22:17.580]Yeah.
[00:22:18.413]Here's, okay.
[00:22:19.246]So here's my thoughts around that.
[00:22:20.640]Yeah.
[00:22:21.810]'Cause you asked, but you didn't.
[00:22:22.770]Yes. But it's fine.
[00:22:27.000]AI is not gonna create it in and of itself.
[00:22:29.820]No.
[00:22:30.659]It'll give you a great pre-draft,
[00:22:33.750]but then you're gonna want to work those ideas
[00:22:37.680]over and over and over until it's good.
[00:22:42.060]So like, there's videos that I'm working on getting ready
[00:22:46.440]for children's literature to film and edit
[00:22:49.200]that it's taken me half a day.
[00:22:53.070]Yes.
[00:22:53.903]For a 20 minute video, to get that script in shape,
[00:22:58.650]in the shape that I want it to,
[00:23:00.060]in the framework that I want it to be in,
[00:23:03.240]and that's, you know, a good four hours
[00:23:05.220]for a 20 minute video.
[00:23:06.750]And so-
[00:23:07.583]That would've taken me eight before.
[00:23:08.640]Yes.
[00:23:09.734]So I will, like- It is a time saver.
[00:23:10.950]But you cannot think, oh, the night before
[00:23:14.400]I need to post it, I'm going to create it.
[00:23:17.105]Oh no.
[00:23:17.949]In five minutes using AI.
[00:23:18.810]That is not going to happen with video.
[00:23:20.916]Nope.
[00:23:21.786]It's not going to happen with the presentation.
[00:23:23.490]It's not gonna happen with anything.
[00:23:25.680]And so, and your level of expertise is critical.
[00:23:29.490]This is what people are coming to see
[00:23:31.980]is your level of expertise.
[00:23:33.900]And if that can be replaced with ChatGPT,
[00:23:36.180]so maybe they need to go to ChatGPT University.
[00:23:38.400]But I argue that for most people,
[00:23:41.700]most of the time, when it's our area of expertise,
[00:23:44.610]ChatGPT, or any other LLM, is not better than us.
[00:23:48.060]Right.
[00:23:48.893]It should be amplifying you.
[00:23:50.340]Yes.
[00:23:51.173]Not the other way around.
[00:23:53.280]Yeah.
[00:23:54.113]So I think that's probably kind of the biggest red flag.
[00:23:58.050]Yes.
[00:23:58.883]Don't let it take over. Don't let it take over.
[00:24:01.500]There was a really interesting study that came out
[00:24:04.350]around people who personified their chatbot.
[00:24:10.320]Yeah.
[00:24:11.160]And how when they're not around ChatGPT,
[00:24:13.920]they miss it, and they become less able to think.
[00:24:17.820]And I was like, ah. (Guy laughing)
[00:24:22.820]Well.
[00:24:25.920]I mean, any habit, right?
[00:24:29.610]Any habit becomes much, much harder
[00:24:31.710]when you can't go back to it.
[00:24:34.200]When we all started Googling,
[00:24:37.860]and finding things online. Oh my gosh.
[00:24:39.420]I was so young. Readily.
[00:24:41.700]it became really, really hard
[00:24:43.680]to start writing academic papers
[00:24:45.840]without being able to just quickly
[00:24:48.600]go Google something, find it,
[00:24:51.450]and like, okay, I'm on the right track or not.
[00:24:54.780]And I think that all of our LLMs will serve that function.
[00:24:58.980]And that is, I always have a window
[00:25:01.470]into one of my favorite chatbots
[00:25:05.550]helping me generate when I get stuck.
[00:25:07.950]Now sometimes I don't use it at all,
[00:25:10.530]but it's always there as an option.
[00:25:12.930]Yeah, I like that.
[00:25:14.330]And so will you miss it?
[00:25:15.600]I don't know.
[00:25:16.433]But I think that if it stopped working, I would miss it.
[00:25:19.961]Yeah, I think if I didn't have access to it.
[00:25:21.660]Yeah.
[00:25:22.575]But I don't like- Miss it, miss it.
[00:25:23.670]It's not like- Oh, yeah.
[00:25:25.354]Oh, I need some GPT. Yeah, that's true.
[00:25:28.320]Yeah, if you're starting to develop feelings
[00:25:30.330]towards your LLMs, maybe it's time
[00:25:33.210]to take a break. (laughing)
[00:25:35.460]Please.
[00:25:36.293]ChatGPT's like, "No, no.
[00:25:37.500]This is why you're single."
[00:25:40.021](Guy laughing)
[00:25:41.850]So that's not even happening.
[00:25:43.673]That's not even possible.
[00:25:45.120]So okay, you've done all this work.
[00:25:47.910]Now oftentimes, particularly online teaching,
[00:25:51.720]gets a, not bad rep,
[00:25:56.820]but gets kind of discounted.
[00:25:58.740]Yes.
[00:26:00.150]There's this weird perception that it's easier,
[00:26:03.240]that it's less work intensive,
[00:26:06.090]to which I say bullocks.
[00:26:08.100]Yeah, definitely.
[00:26:09.900]Because, like, this class doesn't start
[00:26:12.240]until, I don't know, July 5th,
[00:26:14.520]and it is April 2nd, 3rd?
[00:26:17.160]No, 2nd. 2nd.
[00:26:18.180]And I'm already preparing.
[00:26:21.330]Yeah. For that class.
[00:26:22.290]And I'm behind schedule.
[00:26:24.016]And just to give another example,
[00:26:27.300]I was helping a faculty member who needed teaching review.
[00:26:32.700]And this is a faculty member who teaches mostly online,
[00:26:34.950]I think 90% online.
[00:26:36.570]The one time he invited me was to see him teach in person.
[00:26:40.230]And I'm going, you teach 90% online.
[00:26:43.710]I want to see how you teach online.
[00:26:45.690]Right.
[00:26:46.523]Because that's where your expertise is needed,
[00:26:48.810]and that's where you're having most of your impact.
[00:26:50.760]Let's see that.
[00:26:51.593]And that's the discounting, right?
[00:26:53.580]Mm-hmm.
[00:26:54.413]It's real only when it's person.
[00:26:55.440]No, it is not. Mm-mm.
[00:26:57.477]And this ability to teach online is magical,
[00:27:00.660]and as I admitted, maybe a little bit
[00:27:03.870]of a challenge for me sometimes.
[00:27:05.370]So it is a very different thing.
[00:27:07.410]It requires different documentation.
[00:27:09.600]Yes, it does definitely require different documentation.
[00:27:13.372]And I think part of when you're putting up
[00:27:16.890]your tenure file, and you're collecting all your data
[00:27:21.630]to document your work, all of those points of responsiveness
[00:27:26.610]and presence, you want to include those.
[00:27:29.370]Yes.
[00:27:30.203]The thing that I love about children's literature online
[00:27:33.990]is I get outta the way.
[00:27:36.600]Yeah.
[00:27:37.433]It gives my students more time
[00:27:39.090]to actually read the books that are part of the course.
[00:27:42.960]Yeah.
[00:27:43.793]And so for me, that online design works really well,
[00:27:48.060]but that doesn't mean that I disappear.
[00:27:51.510]For every book that they have book choices for
[00:27:54.090]within their genres, I've got a one minute teaser video.
[00:27:58.110]This is what this book is about.
[00:27:59.340]Here's why you might be interested.
[00:28:00.600]So that's a contact point.
[00:28:02.790]Every module I've got,
[00:28:05.070]Dan and I did a bunch of videoing a couple of years ago,
[00:28:09.210]I have videos where they see my face,
[00:28:11.400]and they hear my voice, and I apologize to them for it.
[00:28:15.810]But like making sure those are documented,
[00:28:17.820]'cause that's not nothing.
[00:28:18.750]Those are pedagogical pieces
[00:28:20.970]that when well designed,
[00:28:23.370]and I'm not saying that mine
[00:28:24.203]are like completely well designed,
[00:28:26.520]but like, it's better than if it were just
[00:28:31.230]sitting in front of your computer, hitting screen record,
[00:28:33.600]and going through a PowerPoint slide.
[00:28:35.010]Like- Yeah.
[00:28:35.843]That's work, but that's not the same kind of work.
[00:28:39.090]So you want to really document the innovation,
[00:28:43.200]the thought you put into it.
[00:28:45.000]And then there are different metrics
[00:28:46.740]that you can actually pull.
[00:28:47.790]Because one of the advantages of LMSs
[00:28:51.090]is you can actually go back
[00:28:53.070]and pull how much interaction, how much,
[00:28:56.730]how many people viewed different things, and all of that.
[00:28:59.970]You don't want to do it as part of evaluation
[00:29:02.280]of your students, but you do want to do it
[00:29:04.620]as a way to think about how you showcase your teaching.
[00:29:10.260]Most colleges have, at one point or another,
[00:29:13.950]in their teaching section on P&T document,
[00:29:18.570]innovation in teaching.
[00:29:19.800]This is a great place to show innovation,
[00:29:22.170]especially if you are using AI,
[00:29:23.850]and you want to be very, very clear
[00:29:25.770]about how you're using it.
[00:29:27.420]And we need to start doing that documentation
[00:29:29.280]also because we need to normalize it,
[00:29:31.380]and we need to talk about what we're doing.
[00:29:33.690]I'm really worried about people doing things,
[00:29:36.330]but then hiding it.
[00:29:38.490]Right.
[00:29:39.323]And so it becomes a trap where we're not talking
[00:29:41.730]about what's actually happening,
[00:29:43.290]and we're not evaluating it correctly.
[00:29:45.300]Well, and I think that's one of the interesting things
[00:29:46.710]about an online course.
[00:29:48.110]Yeah.
[00:29:48.987]Is too often measures of your teaching come down
[00:29:51.750]to things like student evals.
[00:29:53.640]But in an online course,
[00:29:55.770]literally your teaching is the most visible thing.
[00:29:58.890]Yes. In that course.
[00:30:00.870]Yes.
[00:30:01.830]So include that, leverage that,
[00:30:04.950]make sure it's an important part of what is happening.
[00:30:08.070]And most of us teach the same courses over time.
[00:30:12.360]So also document the changes you're making
[00:30:15.420]based on the previous feedback.
[00:30:17.310]And that's another way to address
[00:30:19.230]how are you thinking creatively about your teaching,
[00:30:21.900]is this happened here,
[00:30:23.700]and as a response, I made this change.
[00:30:25.860]I've added these videos,
[00:30:27.360]I've thought again about this transition,
[00:30:30.660]and whatever it is.
[00:30:32.130]Yep, nope.
[00:30:32.963]Totally.
[00:30:33.840]All right.
[00:30:34.673]So, in conclusion.
[00:30:37.020]Online teaching isn't about doing less,
[00:30:39.150]but it's about doing things differently,
[00:30:42.240]and making sure that you're doing it
[00:30:43.470]with some kind of strategy, presence,
[00:30:46.230]and maybe living in "The Jetsons" a little,
[00:30:49.380]and getting some help from the bots, but not too much.
[00:30:54.780]And when your AI tools start feeling like
[00:30:58.770]they're your teaching assistants,
[00:31:01.260]or maybe you're their teaching assistant-
[00:31:03.598](Nick laughs)
[00:31:04.431]It's time to pause, pivot, and rehumanize.
[00:31:07.350]You want to make sure that your students know
[00:31:10.020]there's a person on the other side.
[00:31:11.880]So let us know how you are using online digital,
[00:31:15.473]online teaching to work for you,
[00:31:18.390]as we're seeing just this massive shift toward it.
[00:31:21.279]And if you've also had a full AI meltdown,
[00:31:25.290]the doctors are in.
[00:31:26.370]We're listening.
[00:31:28.680]And stay tuned.
[00:31:30.900]Our bonus download for practical tools that you can steal
[00:31:34.890]and start using tomorrow is going to be available soon.
[00:31:38.310]And thanks for being here.
[00:31:40.380]We appreciate you.
[00:31:41.370]And until next time, stay present, stay curious,
[00:31:45.120]and remember, not everything that glitters is AI gold.
[00:31:50.211](upbeat electronic music)

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Not That Kind of Doctor - Online Teaching

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments