Back to the Future of Learning in the AI Era with Michael Wagner

In which Dan chats with Michael Wagner, Professor and Department Head of Digital Media at the Antoinette Westphal College of Media Arts & Design at Drexel University, about the implications for AI in education. 

Back in July Michael posted a piece to his Substack with the provocative title, "The End of Cheating as We Know It." He posits that, yes, AI has the power to disrupt everything in education, but, ironically, it has also created the conditions for ancient practices  of teaching and learning to ascend, methods like socratic and oral examination, as well as design cycle iteration.

Mentioned in the show:
The End of Cheating As We Know It by Michael Wagner
Michael Wagner's Substack
Alpha School

Music: Inspired by Lakey

Transcript

SPEAKER_00 0:09

I would uh recommend that teachers look into kind of how people used to teach a long time ago, because that is really what we need to do. So ironically, what AI is doing is it's bringing us back to the core values of education, right? It's no longer kind of uh output and and kind of creating a product and kind of creating a text and then kind of uh making sure that the text is it's it's really kind of the thought process, the humanistic element of of education that that's starting to count.

SPEAKER_01 0:52

Welcome to What's the Big Idea? I'm your host, Dan Carney. Artificial intelligence is being met in education with the same emotions as in many industries. Fear, excitement, bewilderment. It's the topic of faculty meetings, it's the stuff of conversations in classrooms, it is some teachers changing everything about how they teach, while others are vehemently resisting. Is AI about to change everything, including education? Initiatives like Alpha Schools, which have essentially replaced teacher instruction with chatbots, sure think so. While others, like Friend of the Punt Gary Steger, see AI as another tool among many. Today's guest, Professor Michael Wagner, thinks AI just might change things, but maybe not in the way you're thinking.

SPEAKER_00 1:43

So at the moment, I'm the department head and professor in digital media at Drexel University. I'm in a design college, and I'm sort of the technology educator in the design college. That's usually the way I frame myself.

SPEAKER_01 1:55

Back in July, Michael wrote a piece on his Substack with a provocative title, The End of Cheating as We Know It. He posits that yes, AI has the power to disrupt everything in education, but ironically, it has also created the conditions for ancient practices of teaching and learning to ascend, methods like Socratic and oral examination, as well as design cycle iteration. He also stresses the importance of making learning local and personal in the age of AI. It's an excellent read, and I'll post it to the show notes. I started off by asking Michael about the state of play back when he began his career.

SPEAKER_00 2:34

So so when when I got into the field, it was actually pre-uh computer era. So it was actually in the 80s. And uh during that time, um things were completely different. So so I my my original my original um education is actually as a mathematics teacher with a specialty in what we called descriptive geometry. So this was a a topic that uh was related to the use of uh geometry for communicating architectural drawings or machine drawings. And that was a school subject back in the days. Over the years, I've encountered that multiple times that new technologies uh enter the field and question everything that you do in education. That was once again when computer design came and took away everything that we actually kind of trained for. Uh, and then uh the internet, obviously, um the communication devices in the in the early 2000s. Um, I was one of the first to create a um an e-learning course in the in the 90s. Uh, you know, at that time I was at Arizona State University, and kind of um, and and kind of that that brings us pretty much to the common to the current day where we have the same uh situation again, uh, with uh artificial intelligence questioning a lot of things and making a lot of things that we do obsolete. And it's really interesting to see how that how that progressed.

SPEAKER_01 3:59

Originally, I heard about your work from a substack you wrote over the summer. I think it was over the summer, uh, with one of the most provocative and I think for educator interesting titles I've seen in a long time The End of Cheating as We Know It. And it it really crystallized for me a lot of the things that I've been thinking about and talking about with other educators, about the nature of AI in schools. And actually, just earlier this week, I was in a meeting in which my school was having a conversation about which AI detector we should use. Which AI detector did we want to pay for? Uh, we were going through sort of the merits and and statistics, often self-reported statistics from these companies. But I think, you know, you I think would place this conversation under what you call the collapse of a flawed system. Um, what did you mean by that? And what's your what are your thoughts on AI detection?

SPEAKER_00 5:04

Um my thoughts on AI detection are um well, how should I put this? Well, essentially, I think um we are we are chasing or we are trying to solve the wrong problem, right? Um when we think about AI detection, we think we can kind of uh answer the challenge of artificial intelligence in the same way we answer the challenge of the internet, right? With the internet, we could kind of distribute uh all kinds of papers and people could simply copy and paste and and and kind of re-reuse uh content from somewhere else. And with that came sort of the idea that we can simply kind of look into the internet and see if something that has been written is already out there. Um, what struck me is that with artificial intelligence, the um discussion has to be done completely differently because the system itself does not really replicate word by word. It it kind of recreates something that might have a similar tone or similar content, but it doesn't really uh copy and and and and it doesn't really do any copy and paste. Um so, in that sense, um I started to really question the idea of uh of going out and seeing if something has been uh copied from somewhere else, because it doesn't really make a whole lot of sense anymore. Um, what we're really doing if we're doing that is we are we are kind of uh filtering out the students who are who are not capable of using the technology, right? Uh that's really what we are doing. If if if you have a student who is really confident in using artificial intelligence, they can actually use artificial intelligence in such a way that uh that can bypass any protection whatsoever. There's absolutely no protection or no detector that uh that is capable of uh of catching everything. So, what we are really doing is we are educating our students to become better cheaters by ironically trying to find out who are the ones who are cheating, right? Um, so in in that sense, my feeling was that uh the discussion is mood. Um we we can't we we need to accept that we can't uh detect everything, um, and we need to therefore change our way how we approach education in general, because cheating doesn't make it's it's no longer an something that I as I as an educator can actually actively prevent, right? Um, so so in that sense, um the question becomes more how do we change the educational system to deal with the fact that we are no longer able to identify whether something was produced by a student or not?

SPEAKER_01 7:48

Yeah, and I want to get to that because I think you really capture this idea that schools need to rethink everything. Um, because schools are naturally strapped for time, strapped for cash, strapped for brain space, and and we're reactive. And right now there's a lot of reaction to the AI. What do you just quickly on the detection thing? What do you make though of all these companies that with their self-reported, their 99% accuracy, and they're you know, we have virtually no false positives. What do you make when they make these claims? What do you think make of that?

SPEAKER_00 8:24

Uh well, I mean, I I do understand why they do that, because they are really uh fighting for survival, right? Because uh plagiarism as such is no longer a thing that we really actively pursue because you know no student is really copying kind of word by word. So they need to figure out something else, and the only thing they can do is uh to kind of use that um that EI detection, and they need to have these high levels of detection because otherwise these tools become unusable as well. Uh, because what it really does is it changes the um the requirement who is actually um in instead of instead of having certainty about a certain teaching, suddenly you're accusing a student of something and and there's no certainty anymore, right? Um so so that kind of changes completely um the the who essentially needs to prove what and what needs to kind of be and there needs to be a grievance process, and kind of students need to kind of me be able to have some sort of process that can kind of challenge what the teacher said, and and uh you know, kind of all these things, whereas where in traditional plagiarism it was very clear, right? So here's the sentence, and here's the way you wrote it, it's exactly the same, right? Uh so these companies need to have this extremely high um detection rates uh in order to be able to survive. And my feeling is that they quite often inflate them, um and uh and and create create numbers that that uh institutions feel comfortable with. But I honestly don't see that in reality. Uh now it might be that uh that currently certain systems have this this level of detection, but as soon as a system comes out, it doesn't take particularly long, and people find a way around that. It's actually really easy to do to bypass IEI detection. Um it's and ironically, you can use AI for that. Um but but uh uh it's it's uh it's one of those things that I think the money that institutions spend on these systems is better spent elsewhere, right? Um I do fully uh understand why uh institutions would like to do that. I I emphasize with with teachers who feel uh threatened and and like to have some sort of detection system, but it's not a um it's not a way uh that leads us into the future because uh it's it's a it's a it's a huge game of whack a mo. Uh it's it's uh it's as soon as something comes out that detects something, somebody else finds a way around that. And and really what you're doing is you're just uh filtering out the students who are not capable of using AI correctly.

SPEAKER_01 10:55

Yeah, as you point out in the substack, these detections are always one step behind. And and it's in a lot of people.

SPEAKER_00 11:04

They will continue to be one step behind, right? So so um things there's always this talk about the AI bubble and and the bubble burst, and kind of people are oh, it's it's not it's now it's plateauing. I I don't see any plateauing, honestly. Um and that that is something that I think a lot of people are not aware of. It's still the developments are still accelerating. Um and uh and you know, um it's uh it's it's uh a false uh a false way to go to to kind of try to figure out a way to kind of detect the I rather we should we should find ways to integrate that for sure.

SPEAKER_01 11:42

Um but you know you talk about co-authorship, and I think that to me to me that felt more of like a university um way of thinking about students creating writing. Um to talk briefly about that, what you mean by that before we get to authentic assessment, which I think cuts to the heart of what schools really need to be thinking about from K all the way up through university. But when you talk about co-authorship and the idea of who wrote this, uh how has how has that evolved now in the age of AI?

SPEAKER_00 12:17

Um, you know, kind of AI uh for writing really becomes a tool and how to become more efficient. And uh and uh co-authorship really is the idea that uh you're utilizing AI as a tool in order to get your thoughts onto paper. So it's not it's not something that uh where you ask the AI, write me something, um, and and then essentially you take it and and that's it. Uh but rather it is a it's really a design process, it's a it's a highly iterative process. So the way I usually work is that all the ideas that I have come from my own head, right? So they are in my head and they can't they kind of get out, and then I start to um communicate with uh with an AI agent. And um even though people don't like the idea of anthropomorphized, it actually helps if you think of it like a grad student, for example. So I usually think of AI as regression.

SPEAKER_01 13:08

Can I interrupt you just for a second? How do you because I think that I used to think of AI as being more of a thinking shortcut than a writing shortcut? How do you ensure students aren't in that initial sort of ideation, pre-writing thinking phase? How do you ensure that students aren't the AI is not doing too much in that initial phase and it's not really the students' authentic thinking?

SPEAKER_00 13:35

Um, the the the best way to do that, and that that's where it really helps me to be in a design college, right? So our classes are kind of uh built on the idea of design critiques. So um there's a lot of communication going on, right? So so when we uh when we kind of prepare a certain thing that students are supposed to do, a project that the students are doing, we're constantly talking to them, right? Uh I meet with my students on a regular basis, I meet with me with every PhD student for an hour every week. Uh I I kind of uh so essentially the discussion in class class time becomes a discussion kind of time, right? Um, and by by having that immediate interaction uh between the teacher and the student, um you can ensure that the thinking process is there, right? And then once you know that the student has understood what they're doing and what they're talking about, it doesn't really make a whole lot of difference how they actually end up with the writing. Okay.

SPEAKER_01 14:28

Yeah, so that that having that relationship, that communication. Okay, so they have the thoughts. So I'm sorry I interrupted you before, but now the co-authorship, that's where that kind of comes in.

SPEAKER_00 14:38

Yeah. Um so kind of it it really is a design process. So you're really designing a kind of a paper, you're designing a text, uh, that's what you're doing, and that it becomes an iterative process. And what I usually do is I work with multiple AIs simultaneously uh to make sure that that the writing is okay and there's no hallucination going on that I miss for some reason. And and in that process, obviously, you also need to check everything, right? So so um it has become quite honestly relatively rare that the that the good models are hallucinating. It still happens, but it it happens less and less. Uh, but you need to check everything. And um and then essentially, kind of uh at some point, like any design process, you come to the conclusion that the text is done. Uh so it's it's really more the um this this this interaction that you have. And the way I usually like to describe this is uh I usually I like to think of the AI agent as like a grad student that I'm talking to, right? So when when I'm talking, when I'm giving a grad student an assignment, right? So I'm giving them an assignment and they go out and uh two weeks later they come back with something that they wrote or something that they kind of figured out and and something that they built. And then we discuss uh what worked, what did not work. Sometimes the things that come back are kind of good, sometimes they are not good. And the the the very same way I kind of uh interact with an AI, just in a much, much faster way, right? Instead of having like a two-week period where the student goes out and comes back with something, the AI comes back to me like in a couple of minutes at most, right? Um, so so kind of it becomes this iterative process, and um and that that that's sort of what I mean by co-authorship, right? So it's no longer it's no the the tool sort of becomes a more active part in the writing process. Um, and um I I fully acknowledge that, right? So be being in a higher education that makes it easier because our students are kind of already familiar with that process and they they kind of use the design process in other areas as well. Um how to bring that into schools is a completely different story, right? Um, but but but sort of that that is that is what I mean by by uh co-authorship.

SPEAKER_01 16:45

And I do think that trying to bring that into K-12 schools would present quite a few challenges, uh, but which takes me to the latter part of your piece about how almost maybe counterintuitively or ironically, AI could actually be ushering in deeper learning and authentic assessment because of the necessity to get away from kids just relying on AI. Um, I wonder first, maybe what if you could define what authentic assessment means to you? Because that's a term that rings true for a lot of a lot of educators, but it's also a pretty big term. So when you say authentic education or authentic assessment, what do you mean?

SPEAKER_00 17:27

Well, for me, it's it's really kind of more the um the direct interaction with the uh with the student, right? So um I also wrote a lot about the Socratic method and kind of this idea of uh of really interacting with the student um and making the assessment based on that interaction rather than on the uh on the actual um outcome or in the actual artifact that they create. So it's not really the text that counts, the text becomes relatively irrelevant in the end because the text is kind of created through this co-authorship. What's really important is the thinking process. And in order to get at the base of the thinking process, you need to have the direct communication with the student uh in order to figure out how they how they actually how they actually kind of think. So that's really kind of at the base of that idea. Okay.

SPEAKER_01 18:15

And to this point that I think as you say in the piece that um define this. Oh yeah, you say here we can now focus on designing assessments that not only resist simplistic AI delegation, but met but better measure and promote deep learning. And you could you use this phrase situated cognition. So how how can we design assessments that are not just going to say to kids, hey, we're opening the door here for you to just go ask chat and just plug in the answer? What's the what's the what's the solution here?

SPEAKER_00 18:55

Well, this the solution for for us uh in in in higher education is uh is something that we are very comfortable with in my particular field, which is sort of the idea of the design critique, right? Um so um the the um if if you're if you're if you're uh studying anything in design, what the the way assessments are usually done is that uh that the teacher will meet with the students as a group and then they will have an interaction session. So it's it's it's pretty much kind of the way you think it is. So there will be boards where where they have all kinds of uh images and papers and everything, and then the student will sort of uh explain their thinking process, and then the entire class essentially sits sits next to it and they will essentially uh argue against or for it. And and it's really this uh this constant uh questioning of what the student has done that that creates the basis of that assessment, right? Um now that requires that uh that class sizes are small uh because you you cannot do that in big in big settings, and that's why I think um a lot of what we have done over the last decades uh actually need to be reverted. So we we got larger and larger class sizes. Um not quite sure about K-12, but in higher education, you know, people tell me they have 200, 300 students in a classroom, that's no longer going to be work, right? Um so so we we need to have smaller groups, we need to be able to kind of have more direct relationships between the students and the teachers. We need to change the educational system in such a way that it actually enables that. Um and uh and that's really that's that's really that's that's the most that's the most that that's the most straightforward solution to the problem. Now I'm not I'm not ruling out that people uh figure out ways and how to kind of do the same thing in a in a scalable way. That that's something that I that I think is possible. But at the moment, I I think uh if if if uh if a a teacher wants to um make sure that their assessments are AI secure, AI safe, uh they need to integrate a lot of uh direct interaction with the student. So ironically, one thing that happens, and actually I've talked to many writing teachers, uh, that writing uh classes will primarily be uh created by discussion, right? So it's it's it's one of those ironic things that that that it's really not so much what you've written, but how have you written, why did you write what you write? Uh what is the thought process that went into that piece, and uh and that kind of is something that happens in in a in a questioning session, right? So it's it's it's really it's it's really kind of going back to that Socratic method, um that uh that that uh that we need to kind of look into.

SPEAKER_01 21:37

I really love all that because you're what you're saying is you're getting to the heart of the thought, not the product. Exactly. What is going on inside your mind? And and you speak in the substack about local context, personal experience. How can students bring themselves into the work in a way that a a a chat bot could never do? and and and multi-stage collaborative processes.

SPEAKER_00 22:05

Yeah. Yeah. Um, you know, kind of the these are quite frankly, these are skills that that that people uh used to have in education, right? So so it's it's uh teach that this was one of the primary primary teaching methods really for for a very very long time. And and just just to tell a story I I I recently was at a conference that was a little atypical for me because it was a conference about it was really a humanities conference. I'm not a humanities person so it's it's it's something that I rarely go. But um it was in Greece and uh and uh the the conference attendees kind of went through a couple of sites in in Athens and and uh and they are essentially some of the conference and it attendees kind of explained what happened at the places that we visited and they did that in a way that kind of was very um based on the Socratic method. So it was a continuous questioning. And I was stunned by how well that worked and and I kind of at the end I said this is exactly what you need to do when you want to make sure that that that's that there's deep learning in the each in the era of AI. And they were kind of kind of surprised because that's essentially a very old skill that that has almost been lost in some way um and and so I I would I would uh recommend that teachers look into into kind of how people used to teach a long time ago because that is really what we need to do. So ironically what AI is doing is it's bringing us back to the core values of education right it's no longer kind of uh output and and kind of creating a product and kind of creating a text and then kind of uh making sure that the text is it's it's really kind of the thought process the humanistic element of of t of of education that that's starting to count and uh and there are methods that people have used for thousands of years that that kind of do that and and we just need to kind of re-revisit those methods and and and and and and implement them in our daily lives.

SPEAKER_01 23:59

Yeah I love the humanistic element you just brought up I saw a comment once someone posted it was sort of a Orwellian picture of how education might where it might be moving.

SPEAKER_00 24:12

It was essentially like AI creates the assignment AI completes the assignment AI grades the assignment right and uh and and it's maybe in some ways it's just this logical end to you know industrial education right that we started over a hundred years ago and now here we are where it's just output output input output input output and I I think you're and that that doesn't it it it by the way this is what this was something I actually tried uh AI grading and uh you know um then essentially that's exactly what one of my students said essentially if if if you're creating design with AI and we are writing this AI creating this what are we actually doing right um and and so that's why I'm a little for example I'm a little skeptical about the AI grading. So that's something that I kind of think is is is questionable. But but um and as teachers we are really the adults in the room so it's essentially kind of we we need to limit ourselves what we what we use right uh but um but yes uh so sort of this industrial kind of idea of education uh does no longer make sense um it does no longer make sense we need to go back to what education was really meant to be to kind of uh to to make us better people and and to kind of instill kind of thought processes and make sure that people have uh the the critical thinking skills and the um the understanding of how the world works and and and how they function within the world rather than uh teaching them how to create the uh the best output in the most efficient way right uh it's it's it's it's it's ironic uh it's it's sort of it's it's turning it it's it's turning our education system kind of um it it kind of re-emphasizes those old values yeah but it's it's difficult though especially in K-12 because we have been overtaxed in so many ways and AI in some way provides shortcuts that we we kind of know we shouldn't be using but they're right there in front of us yeah that that's true and and kind of I'm I'm not completely oblivious to those those those things i in the previous in a previous uh job I was actually in charge of a teacher's college so I have uh and and uh we educated teachers of all levels from K to whatever right um and I'm I'm I'm very uh I understand sort of um the challenges that that that that you're facing in K-12. What I have seen though is that um I've always seen teachers rise to the challenge. You know kind of uh there are many many highly skilled very passionate uh very capable teachers out there um and uh my recommendation would be to seek those people out and kind of uh kind of see what they are doing and and kind of uh follow their lead to some extent if you if you're not really comfortable with AI but but I I think I think uh you need to have some sort of an understanding that things need to change you cannot you cannot remain at the status quo that that's not going to work do you actually work with your students on how to use AI is there is there explicit um you know technical philosophical you talk about the use of AI with your students um yes uh and uh you know kind of at two different levels I have one PhD student uh Daya Ramesani and and uh we are currently um kind of and her PhD really kind of goes in that direction and we're currently also working on a book that we that will come out soon uh and uh and uh with her I'm I'm really kind of uh investigating all the uh challenges that you have and how actually education in general will move into the into the future in in uh in my other classes which are more design focused classes so I teach uh game design and uh you know kind of at the undergrad and the at the graduate level um it's it's it's it's interesting because those are students that are vehemently opposed to ai so so um so you need to be you need to be a little bit uh careful about how you frame everything because on one hand you need to make sure that they are once they're leaving the university they actually have the ability to find a job and that will require them to be proficient in AI but at the same time you also don't want to overload them because they are kind of they might actually shut down and kind of say no I'm not doing that right so so it's it's sort of it's sort of a very uh kind of difficult balance that we currently have in certain design disciplines where AI is becoming a a threat to some jobs right um but yeah we we we do that um you know kind of uh at all kinds of levels uh I I I taught a class on conf UI I'm not quite sure if you guys know what conf UI is but it's sort of a a system that allows you to create images locally on your own machine um and uh and we needed to teach that because some fashion designers interest interestingly were searching for people who have that skill um so so that's uh that's yes we have that included um uh but but once again they need to be relatively careful we are we are talking about students who are trying to enter disciplines that are seriously threatened by AI right so so we need to be very very careful on how we do that.

SPEAKER_01 29:25

And how do how do your students think about all this? You know we talk about the design process and authentic assessment and they also AI they're interacting with AI outside of the classroom in the classroom what kind of reactions are they giving in in in this bigger question about to AI or not to AI to AI or not to AI.

SPEAKER_00 29:47

You get all kinds of of uh of responses quite frankly um you know last last uh term uh I think it was last term uh time uh moves by so fast uh I taught a class about research methods and and I decided to because I wanted our students to really become proficient in the use of AI. So what I did is I decided to create assignments that actually required you to use AI because otherwise you would have had a hard time to really complete them. So the assignments were kind of really really difficult. And uh I'm not I'm not going to do that anymore. And the reason I'm not going to do that anymore because I had actually two students in that class that absolutely refused to use AI. So what they ended up doing is they spend hours and hours and hours in order to complete it by hand. So you have everything you have you have students that are vehemently opposed at the in higher education quite frankly right so you have students that are vehemently opposed to the use of AI and avoid it at all cost. And then you have those who embrace it. Now the challenge will be that uh that uh you know kind of uh if if students are not willing to use AI they will they will they will have a challenge when they when they try to find a a job because because they they will be outperformed by those students who are willing to use AI because AI is in the end a a an efficiency it it kind of increases efficiency in what you're doing right um so so I'm I'm I'm I'm also a little bit concerned about those students who are not willing uh opening themselves up to the use of AI but but you know kind of you you have at the moment you really have everything and students that have not slid into what you call metacognitive laziness that have had an education that emphasized oral examination or Socratic discussion and deep thinking from where you sit at the university level how will they be better situated for post-college life post-college careers versus maybe students that got to use AI and used it a lot through high school through university what will the what do you how will they function differently in the world do you think well I mean the um I think it's not an either or or question right so so uh what you really need is you need both um and uh you know either either either if if you are kind of doing one or the other you're not going to be successful the ones the students that are going to be successful at are the ones that are able to combine those two right so so they are using the ITols to make them more efficient and effective um and uh produce better results uh but at the same time have this deep thinking and deep learning and and kind of uh approach to to to education so if if uh so so for me kind of the the challenge will really be that both groups if you have students that are kind of uh refusing to use AI and just engaging in the thinking process they're not going to be productive enough for industry to actually pick them up and at the same time those students who are kind of uh avoiding the term cheating but sort of kind of getting their way through college by just using AI in order to complete the assignments and not thinking about what actually happened they're not going to be able to find a job either you you could you need to you need to have both you need to be able to fully understand what they're doing and at the same time also use AI in order to increase your productivity. I've seen that in some of my classes um when uh you know kind of uh one one thing that became very apparent over the last couple of months really over the last year I would say is that there are certain students when they are able to combine those two things they can actually produce outputs that I haven't seen before. Like for example I have a game design class that uh in in our grad program that is for students that are not necessarily with a coming with a game design background and and usually kind of they they they made little nice little games and kind of you know it's it it it looked like a student game. But over the last year I I got students who essentially went from zero to 100 within a couple of weeks because they were fully engaged in in how the process actually works. How do I create a game how do I sign a game how do I do and then at the same time used AI in order to kind of do that in a very short amount of time. And the result was that some of the results that some of the games that they produced were almost at a commercial level even though they haven't done anything in game design before. But that required them to really do both uh really engage with the topic but at the same time also use AI in order to get the up the product done as fast as possible.

SPEAKER_01 34:23

It's really fascinating and and I education discussions around AI tend to be quite polarized. You do have a lot of educators saying never I refuse to even use this and then other people saying no you have to you have to use it. And I think what you're sort of advocating is this really elegant middle ground where we're developing the deep thinking but we can't ignore the reality of the technical world that they're graduating into exactly so essentially you need to you need to find a middle ground there.

SPEAKER_00 34:57

And and I think for for me the way I see it is really an opportunity to to kind of do something uh substantial and to really kind of make education more kind of uh better in in some ways. Now I would say though is that that uh from an institutional perspective right so if you if you're managing institution um uh it does make sense to have a mix of teachers right so so I kind of I'm a department head so I have like a department where I have uh all kinds of faculty some kind of are full embracing AI some kind of are on the I'm going to never touch that um it is it is of value if students are exposed to both worlds right so I I as a as a as a department head I don't mind to have people in there that say I'm not going to touch that because because of this and this and this because it it provides students with an additional perspective. I think it's also dangerous to kind of uh streamline everything to the same kind of way of thinking I think there's a there's a difference. But in the end I think the um for for the students to be most um to to be most successful uh we need we need sort of a mixed approach we need sort of a balanced approach that kind of uh uses ai for what it's good for and then at the same time makes sure that people that students are still kind of uh thinking uh critically and are your faculty confident with AI themselves because I think a lot of in the K-12 realm certainly a lot of teachers maybe even those that are interested in using AI in different ways because it feels like it changes so fast and I think how do you know how do you feel confident using it and I think that's actually a roadblock to a lot of teachers how can I do this with my students if I myself don't feel confident that it that is quite frankly that is the million dollar question. It changes so fast. It changes so fast. It changes on a weekly basis. When I created that that class I was talking about before uh the config UI class right so I actually had the syllabus set up everything and and every time every every time every week I I needed to rewrite everything because things had changed at that point. So it it it becomes almost like a real-time kind of process the the this you certainly can't kind of create a syllabus and then kind of use that for a couple of years I couldn't even use it for a week I had I needed to rewrite it constantly um and and there are many there are many people uh many teachers that are uh not able or not willing or don't have the time quite frankly you know kind of time a big challenge to really follow that and uh and that development and that is that is a that is a huge challenge um I've I've talked to many colleagues in my university who said well I I simply didn't have time to kind of touch it ever and now that I'm touching it I'm suddenly realizing that it is able to do that and that's actually really cool can you teach me how to do that um so so it's I think there's a certain danger of missing the um of of missing of missing the boat right um that there there's there's uh and and at the same time I fully understand and I quite frankly the same thing for me as well um you know kind of the it it happens to me I'm I'm almost having that as a full-time job now to really just follow up what's happening in the eye right that's my that's pretty much my full-time job and even I miss developments once in a while there are certain things that I completely missed um because things are moving so fast it's almost on a daily basis that something new happens and um and yeah uh I I kind of I I I really see that as a as a huge challenge. Now what I would what I would tell what I would tell people is to rely on the knowledge that their students bring. Now this is this is a little bit different depending on the on the level that you're teaching right so the younger the students are the the kind of the more that uh but but I think at least in in uh starting at middle school up to high school um I I always found it useful and this is cut quite frankly is also something that we taught our our teachers when I was in when I was running the teachers college um and and game design became a big thing right so so learning with games right so I I told them well I mean your your students are are playing games use their knowledge on what game what the games are and how to actually implement them and and you can put that into context uh that's your job as a teacher but you don't need to you it's perfectly fine to simply have the expertise of your students and integrate that into your planning process. I would recommend just uh just uh being on cognizant of what your students are doing and and trying to figure out where things are moving and just being interested in how things develop. The main um danger that I see is that that people are not the teachers are completely shutting off and said I I I'm not going to touch that ever. And that that is a problem because if you're shutting yourself out you also don't know what the limitations are. I cannot tell you how many how many teachers I've talked to who said um I can confidently um identify any writing that a student brings me that is AI generated. And um I usually have to tell them no you can't it's not you can you can you can only identify the crappy ones. The students that can't do it very well.

SPEAKER_01 40:28

Yeah.

SPEAKER_00 40:29

The students who can do that you you can't it's just no way to do that. And um and uh yeah yeah it's it's it's uh it's so so you need to be interested really that's that's all it is right so being interested in in how this uh technology develops and and uh implementing it thoughtfully wow well I really want to thank you for taking the time to talk today um I'd want to just end on this um not normally in the business of predictions and maybe trying to predict anything in AI is futile but if we were looked ahead even five years three to five years where do you think we might be in terms of how students are using it how teachers are using it how we're how what the nature of school even looks like I'm sorry if that's an unfair question. No it's it's not an unfair question because it's a question that really is at the base of my of my job as a department ed right so if you think about it I I get uh graded I get students into into a higher education and uh and our programs are five year programs. So I need to uh design a curriculum in such a way that they actually have a job five years in the future right so kind of this five years ahead thinking is something that um that is is it's not an unfair question. It's a question that I have to deal with on a daily basis. The unfortunate uh answer that I have to give you I have no clue um the things are moving in such a fast pace that uh that anything is really possible and and uh and you know kind of if I would know how things would be in five years um you know kind of I I I would probably be very wealthy. And uh and uh and that's not the case. So so it's it's uh it's uh it's it's very very difficult to predict. It's also very difficult to predict that we quite frankly don't fully understand how the technology really functions right so I mean we do know how it functions in principle. So it's it's sort of a network and but why it works is is not completely clear. And and uh the the the reason and the fact that we don't really know why it really works the way the as well as it does um also makes it really really difficult to figure out where the limits are right so a lot of people are talking about uh currently about sentient AI and everything and and kind of there's a certain understanding that the current systems are obviously not sentient but at the same time we neither know how our own cognition works in detail nor do we know how AI cognition works in detail um and the the the the lack of that knowledge makes it really difficult to predict anything because because you can essentially predict anything. If if if it would not surprise me if like in two years we have a self a self-aware AI it would not surprise me. It would also not surprise me if in two years everything kind of burst and we are back to the way it was like a couple of years ago. So it is a very fair question but it is I think it's completely impossible to answer right now. And that's something a lot of teachers educators and also educational institutions are struggling with.

SPEAKER_01 43:39

Well we'll just have to get you back on the show in two years and see where we are. Michael thanks so much I really appreciate it love this upstack I love all your thinking about this it's uh really helped me a lot and I appreciate you taking the time to talk well thank thanks uh thanks so much for having me and uh anytime uh just hit me up and and uh we can talk anytime a huge thanks to Professor Michael Wagner for joining me to talk about artificial intelligence so many uh important points he brings up it was so funny to me when he mentioned that it's hard for him to keep up with all the changes in AI and that's his job essentially so how could educators or really anyone who's not fully dedicated to that uh hope to to keep up with all the changes and that's what I I would caution any educator out there to uh experiment with AI on your own, try it out, see what it can do, what it can't do. But more importantly have that conversation with your colleagues, with your administration about what do we believe about assessment and what do we believe about how students can show what they understand. What do we believe about the thinking process more than the writing process more than the the processes and methods, the scientific processes what do we believe about the thinking process in our students that to me is a much more vital question than how are we going to prevent AI from being used to cheat? How can we uh stop kids from relying on AI or the other side how can we run headlong into AI and implement it in everything we do? More importantly what do we believe about how students think thanks for listening everyone really appreciate it. Stay safe out there and I hope your school year is off to a great start