Subscribe
Share
Share
Embed
The Transform your Teaching podcast is a service of the Center for Teaching and Learning at Cedarville University in Cedarville, Ohio. Join Dr. Rob McDole and Dr. Jared Pyles as they seek to inspire higher education faculty to adopt innovative teaching and learning practices.
This is the Transform Your Teaching Podcast. The Transform Your Teaching Podcast is a service of the Center for Teaching and Learning at Cedarville University in Cedarville, Ohio.
Jared:This is the transform your teaching podcast. We are on the campus of Cedarville University. My name is Jared Piles, and across from me is Dr. Rob McDole.
Rob:Hello.
Jared:Hello, sir.
Jared:We're doing something a little different. Yes. We are going to play clips today from our series from a year ago that we titled education in the age of AI, and we decided to listen to clips. This is like, remember those eighties and nineties sitcoms that would have clip shows?
Rob:Mhmm.
Jared:Right? And you never expected it when you were ready to watch a new episode as a kid and you sat down and you're like, it's a clip show. Well, ours is gonna be better because we're gonna play clips and react and see whether or not what we said a year ago
Rob:Was it accurate or not?
Jared:Was accurate or valid.
Rob:Mhmm.
Jared:Right? Or does it still make sense or has the technology changed that much? So I have I have eight clips. We may get through them all, we may not, which is totally fine. We'll just play them.
Jared:I'll play the clip, and then we'll talk about it.
Rob:Roll the tape.
Jared:And then when I feel like we're done, we'll or when you feel like we're done, we'll move the next one. Alright. You ready?
Rob:Go.
Jared:This clip is twenty eight seconds long. So here we go. But now for the first time, it's really kind of become a forefront and right in the middle. I mean, right in front of us.
Jared:And I don't think it's going to go away. It's not a fad that's going to. No. And I could if I eat my words and in six months we're not talking about this anymore, I will gladly eat my hat and go away forever, you know? But I don't see this going away, and I think it's crucial, absolutely crucial, that educators in higher education take a stance.
Jared:Okay.
Rob:So first first off, I think it's important for our listeners to understand that I have never seen you wear a hat. So I'm not sure how you would eat your hat even if
Jared:I do wear hats, but I don't think that's the point of the clip.
Rob:Well, you said I'm happy I'm happy with my hat. Yeah. Because that's not gonna happen. Right?
Jared:It's not.
Rob:It hasn't gone away. Mhmm. And if anything, we've got probably chat GPT five staring at us. We're recording this mid August.
Jared:Yep.
Rob:And by the time this airs, there may already be an announcement of ChatGPT five. Which is crazy. Yeah. And that's gonna be kinda scary because it's currently being integrated into the macOS, iOS Yep. IPadOS, all the OSs for Mac and for Apple.
Rob:So currently, we're as of this recording, we're in developer beta two, eight dot one version two. And I already have access to some of of the AI features inside of iOS and macOS. But all I'm saying is it has not gone away.
Jared:I also wish I could say that the need to adapt to it has gone away. But I feel like we are still having the conversations that we had a year ago about it's not going away, and you need to adapt to it.
Rob:Well, think it's gotten worse.
Jared:You think so?
Rob:Well, there's recently been some polls done by Gallup, and 70% of faculty higher education faculty, and I think K-12 as well Mhmm. Who were surveyed, 70% of them say it's a real problem because of cheating. So even with faculty, it's still a big issue, and a majority of them still don't feel prepared.
Jared:Right. And I've, you know, I've been to several conferences since we did this and the overall, perception I get from speakers and attendees is or the speakers are for it. The attendees are like, I don't know what to do with it, like you said. Like, I I it terrifies me. I don't know what to do.
Jared:So there's there's a lot of work that still needs to be done about using it and using it effectively.
Rob:Yeah. The the main thing I've told faculty when we've had those kinds of conversations is you gotta use it.
Jared:Yeah.
Rob:I I don't know how else to tell you that. You just have to use it.
Jared:Yeah. Alright. Clip two?
Rob:Yeah. Let's go. This
Rob:ChatGPT four has already evolved.
Jared:Let's stop real quick because this is when you had your horrible chest infection, and you were this is great. This is this is around the time of Rob is sounding really old or really sick. So here we go. Ready?
Rob:This ChatGPT four has already evolved into something more. They've added more to it. So it can actually browse the current web. Yeah. So you could turn that on.
Rob:You can also turn on a whole host of plugins with ChatGPT four now.
Jared:I have a question. Was this you twenty years in the future when you are old and can barely talk?
Rob:No. That was me coming off of a pretty good sickness, if I recall.
Jared:Yeah. You were a trooper for
Rob:I felt like felt like an asthmatic, and no disrespect to asthmatics.
Jared:Yeah. You're looking at one right across
Rob:from me. Yeah. Well Thanks. And I think I was actually at the time too because the doctor gave me, was it albuterol?
Jared:Yeah. Yeah. So I had to Good old steroid.
Rob:Yeah. I was taking that too, but it also sound like it was about to die.
Jared:Yeah. I honestly don't remember what you said.
Rob:Well, I think we kinda I think we kinda covered in the first one. Like, things have been added. Right? And and it's access to data. But I think the other thing too that I've seen is that the government and others, even bodies out there and and that deal with standards with data, you know, they're bringing a lot more control.
Rob:They wanna bring a lot more control to what data AI has access to and how and all those processes. So I think we're gonna see more of that. I think we're gonna see more standardization. Mhmm. Whether or not they can keep up with it, that's a whole another that's a whole another ball of wax.
Rob:And I know OpenAI has lost some leadership along that lines where they were there for, you know, safety and making sure that things were being run properly. And this is no admission that things aren't, but a couple of them ended up going to Anthropic Claude because I think they felt like probably the way things were headed at OpenAI maybe is pushing it further than they felt comfortable with.
Jared:Gotcha. Alright. Ready for the next clip?
Rob:Yeah.
Jared:Let's do it.
Rob:Really, the only people that can have access to these kinds of AI models right now are places like Microsoft, Amazon, Apple, anyone who has a very, very large cloud computing infrastructure because it takes a lot of hardware, you know, to to make this thing run. Right.
Jared:Well, that's no longer the case.
Rob:Oh, it still is. I mean, we we have we have schools that are doing it with open source models, but, like, I think it's University of California San Diego, I believe.
Jared:Maybe I misunderstood the clip. I thought it was talking about access to ChatGPT or OpenAI.
Rob:Oh, running it.
Jared:Running AI.
Rob:Running AI.
Jared:Okay. Gotcha.
Rob:Yeah. Because the reason for that comment was, basically, if you don't own the data
Jared:Mhmm.
Rob:You can't control what it's gonna say.
Jared:Gotcha.
Rob:And these companies are controlling that.
Jared:Right.
Rob:I mean, whatever just like computing. Garbage in, garbage out. Good data in, good data out.
Jared:Mhmm.
Rob:So the algorithms that are running are running on the data.
Rob:And if it's bad data, it's gonna be you know, it's not gonna give you the kind of results you want. Yeah. So my insight was we need to have something that we're controlling ourselves. We need to control the data, and we need to control the model. And the problem is the amount of power needed and the computing power needed is beyond what we can afford as, say, like, just here at Cedarville.
Rob:Right. So University of California is running an open source model. They spent, like, I wanna say, $55,000,000, somewhere in there, a large amount of money to stack something up. Mhmm. At least that's what my sources tell me.
Rob:Mhmm. And they're offering this to other institutions not too dissimilarly from what we pay for API access to OpenAI or other types of models like that.
Jared:Right. Right.
Rob:But cheaper, a whole lot cheaper, but still, you know Yeah.
Jared:It's all about the control.
Rob:Control of the data. Who's who's controlling the data? Who's who's putting guardrails around it? What kind of guardrails are they putting around it? What is it allowed to say?
Rob:What is it not?
Jared:Reminds me of the book I'm the second book I'm currently reading because I decided to read two. Brave New Words that, doctor Jody Penrod recommended. By Sal Khan. It's by Sal Khan, the creator, CEO. Is that his title of Khan Academy, something Yeah.
Jared:Like he talks about them creating their little bot, Khanmigo.
Rob:Mhmm.
Jared:And he talks about the guardrails and fences they put on it so that false information wasn't like, used the example of round earth versus flat earth. Right. And whenever you were like, I wanna argue against round earth. Round earth, is that the right term
Rob:you used?
Jared:I feel stupid for saying this, but, and he was and the the bot refused to interact. Correct. Which, you know, cool, but also in the wrong hands, that could be another propaganda machine that you push out.
Rob:It's like who determines Yeah. Who determines what's correct? And it and and the answer to that is whoever has control over it. Correct. Which is whoever has the money.
Jared:Next clip. Let's go. Clip four.
Rob:So it's kinda hard to say where things are. It's kinda gotten a little secretive here. Recently, once Microsoft put in billions of dollars in in funding to OpenAI.
Jared:Well, they had something called Copilot that they were gonna put in Word.
Rob:Right? It's there. Some companies have it.
Jared:Oh.
Rob:And they're using it.
Rob:It just hasn't hit the mainstream yet.
Jared:Well
Rob:Yeah. It's hit mainstream.
Jared:It hit it's hit mainstream by now.
Rob:Yeah. It's it's really not that good. Yeah. It's true. We we played with it.
Rob:We all got Copilot, and we started trying to use it. And now maybe it's different now, but it's like the experience we had was such that we're like, okay. I'm I'm tired.
Jared:I'm done.
Rob:I'm done. I don't want to go back in and have Copilot say to me, I'm sorry. I can't do that. I mean, how many times has it told me inside of Microsoft Word, I'm sorry. I can't do that.
Jared:Or just inaccurate information.
Rob:Yeah. You know?
Jared:It's like, why? Okay. That doesn't make any sense.
Rob:Yeah.
Jared:But maybe over time, if we do this reaction to this episode a year from now, maybe Copilot.
Rob:Think I mean, it'll have to be. Yeah. It'll have to be. Because
Jared:I would hope so.
Rob:I mean, what what Apple's doing and and others I mean, others are gonna respond. Models are gonna respond. And, you know, some some things that I've been listening to and reading that are out there kinda on the front lines of this stuff, you know, it's it's indicating that we're getting really close to a point where, yes, OpenAI will continue to be the leader. That's kind of that's kind of been undisputed, especially in in the realm of business and the projects that are going on in business. Everybody kind of says the same thing.
Rob:But the other thing that they're also saying is that the small open source models have gotten to a point where for many use cases, they're good enough. They do the job, and they cost a whole lot less. And so there's a sense in which I think you're gonna have a a fragmentation of the market much like we have cars. Right? You know, Cadillac, GMC, Buick, Chevy.
Rob:So you have this, you know, categorization of of a particular market. I think we're gonna have the same thing with with AI bots.
Jared:I think that's gonna happen with specific business markets as well because I was talking to my wife who's in finance, and she said there is something called tax GPT that a bunch of programmers, I guess tax guys, I don't know, they put it all together and she can use it as she said it's kind of rudimentary at this point. It's not really that great. But you'll you'll see a lot of maybe market specific
Rob:Yes.
Jared:Things like that. Yeah. Like fantasy football GPT or, hey, Ryan, maybe there's a Reds GPT. I mean, that would be amazing. Right?
Rob:An AI devoted to making the Reds better. Yeah.
Jared:Oh my gosh. They could really use that. All I would say is sell the team. Red's Just sell the team.
Rob:Red's manager GPT.
Jared:Oh my gosh. I would love I honestly think and we'll leave this in. I honestly think that a a GPT bot could do better than what the current
Rob:Oh, wow.
Jared:The current front of house and CEO
Rob:I sure hope we don't get, like, a a requirement for a cease and desist or a retract that statement, sir.
Jared:I would be floored if they were listening. That would be amazing, but also I have no regrets. So that's my official statement on it. I'm sticking to it. We have a press conference.
Jared:All I'll say is I have no regrets. Alright. Next clip. Let's go. Let me fire it up here.
Jared:Here we go.
Rob:But the one that sticks out to me is Khan Academy's work with AI.
Jared:Yeah. Mhmm.
Rob:And and how it's developed over time. And it it's pretty amazing. It's you know, if you get a chance and anybody can go and do that, not just your kids, but you could go
Rob:Yeah. And and use their their AI system.
Jared:Khan Academy.
Rob:There you go.
Jared:Khanmigo. Khanmigo. You
Rob:speak to this one because you've been you've been listening to the book here recently. I've already listened to it.
Jared:It's it's spectacular. Yeah. It's it's it's impressive what you can do. It it made me want to we we did a coffee drop about this a while back with Jason and doctor Ye. Yeah.
Jared:And we talked about Conmigo and the impact it would have in education. And it this this book, I'm about two thirds of the way through it. It makes me want to develop our own Yeah. Cedar migo Cedar beets. Or something like that.
Jared:Cedar bot. Stinger Bot. Will allow us to tap into the resources that we have on our campus and set up those guardrails like he had. I mean, it's incredible, the stuff that they've programmed it to do.
Rob:Yeah.
Jared:And it doesn't require knowledge of programming languages to do it because you can just have a normal conversation and say, here are your guardrails, here are your fences, which is super appealing to anybody. Right. As long as you know what you're doing, you can have it do anything you want it to. And I'm kind of hoping haven't finished it, but I'm hoping that he reveals a bit of how they programmed it. He may not, because that could be proprietary stuff and you could have a patent on all that stuff or whatever.
Jared:But in reality, you could do it yourself as long as you have a team of however many people he has, and you could do it in record time like they did. I mean, it's just a powerful tool and it's just it's inspiring. I appreciate his approach to it. I brought it with me and I had to Emily and I went away and, some people saw the book that I was reading. We had whole conversations about using it and the chief concerns were, aren't they going to use that to cheat and all this other stuff.
Jared:And so, but the way he addresses it is that, yeah, they could, yeah, but if you address it, how we've done it and you see it as a tool and as a way of writing with a student instead of writing for the student, Mhmm. Helping them study. Yep. It's incredibly powerful tool. And I I strongly recommend that anyone seek it out and try it out and see what it does.
Jared:It's it's it's remarkable. It's incredible.
Rob:And it is free to teachers.
Jared:It is free to teachers. Yes. Yeah. You'd be crazy not to use
Rob:it. It's I think the thing I like about it the most is the Socratic nature. Yes. So it never gives the answer directly to a student. It always leads them down the path towards that answer.
Rob:And, again, I don't I don't think you're wrong. I think at some point, we've gotta get to a point where we've we're running these these models ourselves. And we're also having conversations with students on what do you do when you go to work for a company and they have their own model and it it kinda dictates, you know
Jared:Mhmm.
Rob:Some of those things. That's gonna be an interesting ethical dilemma for our students.
Jared:Alright. This clip is only nine seconds long, so I have no clue what this is.
Rob:Oh, goodness.
Jared:We'll see what happens here. So be ready to respond.
Rob:Alright. I I I do think Turnitin and others are gonna have a hard time eventually figuring out what's AI and what's human.
Jared:You wanna take this one that you said it?
Rob:Well, I I think it's still true. I mean, I don't think we're any further down the line with that than we have before. Mm-mm. It feels like a cold war
Jared:of AI. The cold war of AI. That's a book.
Rob:Well, you know, at least with this, oh, we found a way to find you. Oh, no. We found a way to mask it.
Jared:Right.
Rob:We just keep going back and forth, back and forth, watching a tennis match essentially, and nobody seems to be able to win a point.
Jared:Sure.
Rob:So it really comes down to design in the classroom and teaching people to use it responsibly.
Jared:Yeah.
Rob:And that's that's a philosophical issue. That is not, you know, that's an ethics issue. That is not Sure. Something I think that can be solved with software.
Jared:I mean, in a real world scenario, I think you wouldn't have to worry about Turnitin checking your students work for, well, plagiarism and for generative AI usage. But the in reality, you're gonna come across it. So Right. It's still an unanswered question. It's just still something that struggling with.
Jared:Next clip. Ready?
Rob:Let's go. Yeah. What's really interesting with ChatGPT right now as of July 20, their just most recent release is that you can actually put in custom prompts or custom statements that will be there every time you run an inquiry. So basically, you can create a profile if you will, that it will keep in the background and use in the prompts that you don't have to keep retyping certain things.
Jared:This is by far the favorite my one of my favorite parts of using generative AI and JADGPT is the fact that you can create, you start it off, the whole conversation with, I always started with, you are this. I need you to do this. I'm acting like this. Do you understand? And it, I use it when I did my dissertation stuff, I had it help me with the data analysis.
Jared:I didn't have it write it for me. Yeah. What I did instead was say, here's my data. Here's what I found. What am I missing?
Jared:And it was like, oh, you considered this and like and it was instead of me hiring a writing tutor
Rob:Mhmm.
Jared:Or someone who's an expert in quantitative data
Rob:Right.
Jared:I had Chad GPT do it for me, and I saved a ton of money
Rob:Which you would.
Jared:By switching to GEICO Which but also using Chad GPT. But, like, it it was a huge because it was like, well, you could look at this data, look at this because, you know, my I have very limited knowledge of quantitative data right now at this point. I took two classes on it and even then, I mean, come on. It's a very diverse, complicated, especially when you get into all the different ways you can run it. But it allowed me to see it from a different perspective.
Jared:And then I could ask it things like, well, what other tests would you run to find any significance in this data? Because you could try this, this, this and this. So, I mean, it's incredible.
Rob:Yeah.
Jared:But that all started from me starting with a prompt that said, hey, you're a quantitative data expert.
Jared:I'm writing chapter four of my dissertation. I need help analyzing the data. Do you understand? And then boom, we were running. Yeah.
Jared:So alright. We have time for one more clip. And here we go.
Rob:It sounds to me that neither of us are for wholesale adoption. We're middle of the road. We gotta adapt. What are the pieces that we need to adapt, and what are the places where we should stray away? We've both used it quite a bit
Jared:Yes.
Rob:To do different things now.
Jared:This was from the adapt, adopt Abandoned. Adbandoned episode. Adbandoned.
Rob:Yeah. So where are you?
Jared:I am a 100% full. Full send. I do everything. I I've seen the power of it. Actually, I had it help me with this whole entire conversation.
Jared:I have it up on the background.
Rob:Oh, do you remember?
Jared:No. Don't. I was gonna say, Rob said this. What should I say? I'm I'm full send.
Jared:I'm I'm into it. I I don't think we should shy away from it. I think we should try to plug it in wherever we can Mhmm. For remedial tasks or things to, you know, in education alone, I think the power is there to have it revolutionize education for the good and be used as tutors, as TAs, to help plan, to help do anything you need it to. Right.
Jared:I'm full send. I'm all for it.
Rob:Yeah. I would say I'm I'm for it as well. My caveat would be you still need to be an expert in your field. Yep. You cannot let it do your thinking for you.
Rob:You need to see what it's written, see what's generated, and check it with what you know to be true. And then I think on the other side too is, you know, I think there's probably something to be said for being open and honest when you use it, which, you know, you've shown.
Jared:Mhmm.
Rob:You've shown that with your blog post that you did and when you were doing some work on a course. And basically saying, oh, here's how I here's what I put in. Here's what came out of it. Here's how I modified it. Those are I think those are adequate.
Rob:I think they're they're right. I think the challenge with that is, though, we don't do that with human beings. Yeah. You know, with teachers. And how many times have you gone to a teacher and said, I'm having this issue with this particular subject?
Rob:Say, mister l. Right? Sure. You know, when you're here as a student and mister l gives you this guidance, and that guidance changes the way you think about something and it also changes the way you do a particular assignment. But you don't say to anyone, well, mister l said here, and this is how I changed my, you know, my result or changed my thinking.
Jared:K.
Rob:We do that in other ways too, like with writing. How many times have you been in as a teacher, been working with a student, and you've given them ideas, and then they write those ideas into their paper and you see it, but you don't take points off because of the idea that you gave them.
Jared:Right. So That's a good point.
Rob:I mean, I mean, just I'm just saying, I don't know if it's necessarily being what's the word? Consistent?
Jared:Mhmm.
Rob:Like, well, you gotta say if you use ChatGPT. But if you go and ask somebody else, no, you don't need to do that.
Jared:That's interesting. I never thought about that. Now I'm thinking really hard about this and Don't questioning It's it's it's that's a great point, and I don't have time to talk about it because we're out of time. But Well. Maybe we'll bring it up on another episode.
Rob:It's a cliffhanger.
Jared:Maybe it's a coffee drop later down the road. This was fun. I really enjoyed listening to clips of us It was a year ago. Yeah. I say we do this again in, like, six months.
Jared:Yeah. How about that? Let's do it. Respond to our own responses. It'll be a response clipception.
Jared:Response clips. It'll be great. That's gonna do it for us on this episode of the transform your teaching podcast. Be sure to like and follow us on your favorite podcast platform. Connect with us on LinkedIn.
Jared:You'll see a link in our show notes and description on how to do that. Be sure to send us an email at CTLpodcast@Cedarville.edu. And as always, check out our blog, cedarville.edu/focusblog. Thanks for listening.