303 - The Next Level of AI

Have you heard of ChatGPT? Probably so, since this name is out everywhere in the education world now. What started as a mention has now encompassed the entire realm of education. This new AI product can do wonders - from writing you a song, to creating essays, lesson plans, and more. So what exactly is it? How can coaches harness the power to assist their instructional planning and help educators find ways to utilize it in their lessons while working around any worries they have? That is what we’ll look to answer in this episode of Restart Recharge!

Podcast Team
Hosts- Katie  Ritter & Justin Thomas
Editing Team- Michael Roush, Justin Thomas 
Social Media/ Promo Team- Annamarie Rinehart, Lisa Kuhn, Maggie Harris
Creative/Content Team- Justin Thomas
Producers- Justin Thomas

Forward Edge Coaches Camp
An event that assists instructional coaches learn new tips and strategies for implementing their rol

EDU Coach Network
We're former teachers turned instructional coaches. Coaches can change student learning for better

Forward Edge Google Educator Bootcamp
This Google for Education Bootcamp is for educators to get certified in Google products

Disclaimer: This post contains affiliate links. If you make a purchase, I may receive a commission at no extra cost to you.

Transcript

Brooke Conklin: 0:00

Calling all instructional coaches join forward edge coaches camp in summer 2023. Coaches camp is packed with high quality professional development exclusively for you. Attendees will work with like minded coaches on creating strategies for building teacher relationships, executing coaching cycles and building a culture of coaching and tech integration within their district. There are two opportunities to attend coaches camp in the summer of 2023. You can join us virtually June 12 through 14th. Or come visit us in Cincinnati on July 27, and 28th. Please visit forward hyphen edge dotnet slash coach camp to reserve your spot today.

Katie Ritter: 0:41

Aloha, I'm Katie Ritter.

Justin Thomas: 0:43

And I'm Justin Thomas. And this is the restore recharge podcast, a podcast by coaches for coaches. We're bringing the tips and tricks in everyday work as an instructional coach to help you out with that, or whatever they call you in your school district.

Katie Ritter: 0:55

Yes, so hopefully you're gonna leave this episode feeling just a little bit less on your own coaching Island.

Justin Thomas: 1:00

All right. Well, the main question here is have you heard of chat? GPT?

Katie Ritter: 1:04

Who hasn't? Right now, Justin, right. Obviously,

Justin Thomas: 1:08

this is a name that is out everywhere in the education world right now. So what started as a mansion has now encompass the entire realm of education. This new AI or artificial intelligent product can do wonders from writing a song to creating essays, lesson plans and more. But we want to dive in and figure out exactly what is it and how coaches can harness the power to assist their instructional planning and help educators find ways to utilize their lessons while working around any worries that they have. So that's what we're going to do to take a look here and the restart recharge to hopefully get you a little bit of an answer with that. And we have an excellent guest with us here today. We have Matias Holzer, Amma Tez is an instructional design coach currently working here at Ford Edge in Cincinnati, Ohio, where he assists teachers with implementing new strategies, curating edtech tools and helps admin achieve building and district goals. Prior to his current role on the team. He has taught foreign language for 10 years from the kindergarten to university levels and has a served as the president of the Kentucky chapter of Association American Association of Teachers of German. He has also the chance presented multiple World Language conferences as well as the Bluegrass intelligent Colloquium. He holds a Master of Arts and teaching from Northern Kentucky University and Master of Arts from the University, excuse me from Transylvania University in sociology, and then outside of teaching, and Matias spends his time coaching high school soccer or tinkering in his workshop. And it is very important that we mentioned this but but Tez is German.

Matthaeus Huelse: 2:42

Right? The indeed, I forgot myself when I was begging. I

Katie Ritter: 2:47

grew up in Germany until you went to undergrad, right? Yeah,

Matthaeus Huelse: 2:50

exactly. Yeah. Well, we

Katie Ritter: 2:52

are so happy to have you as a first time. First of many I'm sure on the podcast. Mattia, so welcome. I'm

Matthaeus Huelse: 2:58

so excited to be here. This has been something I've been looking forward from the very beginning. Excellent. So

Justin Thomas: 3:02

you're here with us. And you're going to be the one that chatting with us about this, this kind of abstract idea, but it's everywhere today. So we're talking a little bit about G chat GPT. But also, we want to just make sure that for listeners out there, we kind of really understand where that comes from. And it comes from this idea of artificial intelligence or AI. So can you give us just kind of a little bit of an idea of what exactly AI is, because it's something that's probably used a lot more than people might even think.

Matthaeus Huelse: 3:28

So when we think of AI, there are several different ways that we've already encountered, things that might be aI already, they are little pre steps that we might have encountered along the way to getting charged GPD today. So you might have seen, do you remember when you had to log into a website and it gives you a quick human test to see whether you are another robot and it asks you to recognize pictures?

Justin Thomas: 3:52

It means once I always fail? You and me both?

Matthaeus Huelse: 3:56

Like how many where you can see a car versus a fire hydrant and stuff like that, right? That's already been used to teach AI right? And when you pick up your phone and you start writing a text message, does it give you suggestions of what the next word would be?

Katie Ritter: 4:11

My my Gmail does daddy accurately my

Justin Thomas: 4:14

text message do but it's never what I want.

Matthaeus Huelse: 4:17

Those are the early steps of AI. Those are some of the things that AI has already done on the way to becoming chat. GPT okay, but in order to fully understand that, that's it. There are a lot of components and AI can look very different depending on how you're using it and how you taught it to operate. Coming in here. I was gonna suggest if anybody is interested, do you want to learn how to build your own AI? I can give you instructions, five steps, super easy. Give it to us. I feel like if you understand if we understand how to how it operates, we can understand what its limitations are which we think of it as being so human.

Katie Ritter: 4:54

Can I ask the question before you tell us how to build it? So when you say like it and it's Learning is there like this one AI net work that things are pulling from or it's like, I would build my own AI. And it is separate from like a Google search searching all of AI out there.

Matthaeus Huelse: 5:15

So there's two components that you would have to start with, you have to think of what am I going to give it to learn from? What kind of data do I start? Yeah. So oftentimes, it's a step of scraping as much text from the internet as possible possible. That would be books, articles, maybe some human interactions and forums, and those chat, those capture, as I mentioned earlier, and then you would have to start thinking of how is this program that I want to create going to be designed, and you call that a neural network, supposedly, where you have layers, and it is trying to learn by making a mistake.

Katie Ritter: 5:52

But so it's my own neural network that I would be building if I am building my a. So chat. GPT is its own neural network that may be separate from another AI system out

Matthaeus Huelse: 6:04

there. And it lives on a cluster of computers or servers, where they have all of the data stored. And it does a lot of calculations, a lot of the same stuff over and over again really, really fast. Okay. And that's kind of what it does when it learns. So the hardest part of

Katie Ritter: 6:19

my neural network is like grass. Neurons are forming.

Matthaeus Huelse: 6:25

Perfect. That's the hardest part. I think in order to make this not in order to not get into the weeds. The software part the algorithm, that is the hardest part, I say, for this, we just put our witches hats, hats on, put a colon in the middle, we're going to throw in a couple of algorithms, a library functions classes, we mix this all together, we throw in a little thing called tensor. And then after that, we have pretty much that network that is able to learn by making a mistake and adjust its behavior and recognize patterns really, really well.

Katie Ritter: 6:57

Okay, so now I'm picturing Winifred Sarah and Mary Sanderson and Hocus Pocus, mix it up an AI network now

Matthaeus Huelse: 7:07

that is perfect. That is, that is exactly where I want you to go. Yeah. See, that's

Justin Thomas: 7:11

that's good. I always think of when I think of like aI think of with like, Iron Man. And like the Avengers, because that's kind of what Jarvis is. Right? He's AI and he's learning. And then when they implement all Tron, right, he learns very quickly, and then goes, the more destructive route and thing so like, that's kind of how my mind goes to it, maybe more so than the witches. So he kind of looking at me like, What are you talking about right now?

Katie Ritter: 7:35

A little bit just, I think I know what the first Jarvis is.

Matthaeus Huelse: 7:42

Talking about. Appreciate the Marvel reference right here. I'm very happy about

Katie Ritter: 7:46

so what's this tensor? And if it's too in the weeds, because we just want to kind of get an initial understanding. But you said the last like dash of tensor, you said, What is that from a high level

Matthaeus Huelse: 7:55

think of it is as a rule, or a set of rules for the AI to learn from. So when initially, we put the AI in front of that data, and it just tries to recognize patterns. And that's called actually unsupervised learning. Okay, which I think as an ID, Coach, I very much appreciate that term. So it learns to recognize patterns in the data first, and then in the next step, we tried to give it a little bit of instruction. So we tell it to maybe recognize sentences or structures or pictures or certain things in that direction. And that's when we start giving it a couple more instructions. And those libraries, those rules, those functions on that, that give it some more instruction that would be TensorFlow, for example, that is one of those things. Okay, got it. Thank

Justin Thomas: 8:40

you. Okay, so now let's dive a little bit more with chat. GPT. We've talked a little bit about it. But as Katie mentioned, stands on its own neurological network. But what exactly is it? And what are some of the functions that you've seen done with it so far?

Matthaeus Huelse: 8:54

So what it does really well is it recognizes patterns, extremely great, and a great set of data. So we have been slowly but surely scaling that amount of data and finding the ways that fine tune the ways that it has been learning, which has made it impossible, incredibly great at which has made it very good at understanding human speech, understanding the implicit and the implicit parts of our speech where we might have a computer look closer at certain parts of the sentence to figure out what exactly you're looking for in the answer. It has gotten really good at creating sentence creating sentences that sound human to us. And it can synthesize all of that information that has in the background to then create solid guides for us humans to follow and learn from

Katie Ritter: 9:42

it. My my understanding Mathias is that it's really meant to be like dialogue. More so like it's not a Google search, right? Like I tried to ask it to write me a paragraph describing forward edge because I was super curious what it would say a And it couldn't because it couldn't access the internet. And unfortunately, we do not exist and whatever encyclopedia it's been fed to learn. So it didn't know who we were. Is that an accurate statement?

Matthaeus Huelse: 10:15

Absolutely. So chat GPT specifically has been designed to be a assistant in the sense that it can understand what you're asking it to do. And it can do it in a way almost a human would want to do it.

Katie Ritter: 10:27

So give us some examples of like, what you've played around with and what you've seen some other people play around with.

Matthaeus Huelse: 10:32

So the immediate response of almost everybody that I talked to has been to try to have the AI create a text for you, which is super exciting, right? We want to hear it and make make a little like, poem or a little short story. But you keep extending that and you start thinking about what can I ask it to do? I have used it to just learn new skills I have, I remember the first time I ever used it, I wanted to use create, create a Excel spreadsheet, and I didn't know the exact function I needed to use. So I just asked Chad GPT Well, hypothetically, if I want to pull this data from this column, and have it here and do this to it, how would I do that? And it just spit out the correct answer and the code I wanted to use. So I kept playing around with it a little bit more. And I learned that if I give it more information, such as the exact column and the exact name of the sheet, it would create an even better answer, right? So you keep playing with it. And you learn that it can give you far more than just answers and instructions, it can also start taking the info that I gave it, such as meeting notes, I was sitting in the meeting with a principal and I was recording it. And I had a transcript of the meeting. So I then copied it, pasted it and just chatted btw and I asked her to tell me what are my top three action items I need to take? And how should I schedule my next week to do it. And it blocked out today's for my research that I had to do for the exact tasks I needed to do by day in order to reach every single one of my tasks on time. And you keep exploring and you go further and further.

Katie Ritter: 12:04

My head is literally Brain exploding emoji. My neural network can't take much more. That's awesome. I've seen people using it to just to write like emails. Especially I know, we've had some folks on the team practice with like a difficult parent email, let's say or like inspiration for a blog post or, or whatnot. So Mateus I love that you're super advanced with this spreadsheet, I don't know that I can even wrap my head around how to ask it.

Justin Thomas: 12:37

That's what I was seeing you saying I was like, I don't even know where to begin.

Katie Ritter: 12:40

I love it. I'm like trying to get there. My favorite

Matthaeus Huelse: 12:43

part is actually I'm learning how I'm communicating with the API is changing over time. At first, I initially started to feel bad when I asked it to do extra tasks. When I thought like, oh, no, I don't have to do this myself. I don't worry about that. And I felt guilty. But now I'm realizing no, it doesn't care, it'll do the job for me was the exact way I wanted to and I have to worry about it being overwhelmed with my request.

Katie Ritter: 13:06

Oh my gosh, that's funny. So obviously, you know, you're we're talking about how it is conversational, and it writes text and all of these different things that it can do, as you just said very well, without any complaining. So thinking about that in the context of education, obviously. I think when we say AI, you know, just in your I forget the bad guy example. But you know, AI Yeah, Ultron going the wrong direction, right, like AI in general makes people nervous from the learning standpoint. And the Will this be bad, right? Yeah. But I would say as like a sector, I feel like education's got to be the most concerned with this technology. So how might you think about addressing some of those concerns that educators have about students no longer submitting original work and be their original work? Because the text itself is original work, right? It's not copying and spitting out, and it won't give the same two people the same prompt. So the chat GPT text that is spit out is original work. So turnitin.com different, you know, plagiarism reports aren't going to detect this that the student has plagiarized. So education is nervous. Our coaches listening to this know that their teachers are probably freaking out a little bit if they're aware of and have heard about this. So talk about some of those concerns and how we might begin to address some of those.

Matthaeus Huelse: 14:44

I think plagiarism is obviously the first thought that all of us had, as soon as we saw that there's this chat box you can throw in a task and and it'll create a story for you. And while I think that that is is an issue and that is something that we need to worry about. And we need to definitely adjust as we're moving on in education. In the long term, there have been so many different ways that we have been challenged in terms of plagiarism, that we have adjusted to that. I don't think that this is an immediate concern. I mean, we have, we have seen the rise of Google and Wikipedia, get wrote worries up in us and worry about what does that mean for our students, I think what we need to remember is that open AI has a huge incentive to also make sure that whatever data they produce is going to be watermark. So open AI is working on watermarking it. So in the context of wanting to make AI better, we need to have good data and open AI wants to make sure that it can separate whether something on the internet is created by itself, or whether it's created by humans, otherwise, it'll the AI models won't be as good. So open AI is very motivated to make sure that they have an AI checker and a watermark in there. In the long term, there will be ways to recognize AI production as opposed to human production. Do we need to adjust in the long term? In the short term? Absolutely. In the long term, there will be easier ways to do

Katie Ritter: 16:09

when you say long term like I this is not founded in anything other than like the market is there now for something to come up with an address like an AI checker. And so I see that being like, hey, put our emphasis on this because we're gonna make a lot of money if we sell this to schools. Yeah. Um, so I when you say long term, like how long term are you thinking? Because I would think we're going to start seeing something like within a year, yes, I would have to think if not even sooner,

Matthaeus Huelse: 16:37

I agree. And like long term is also relative. This is, I mentioned this over and over again, this is somewhat relative. Somewhat similar to our experiences, when the internet coming up for the first time, there are so many things that we don't even know we need rules for. And we as a human society, have been watching them become a thing. And we had to make rules and adjust rules and change rules. Same thing is going to happen with AI. But when I say long term, but what is so intense to me is, I don't know what long term really means it could literally be just a year, if you compare the amount of users that were accumulated by Spotify, for example, when it just came out, it took three months to get, I don't know the exact number, a certain number, it took only three days to catch up with Spotify in terms of users. Yeah,

Katie Ritter: 17:24

yesterday when I pulled up chat GPT, which on the day, we're recording this, that would have made it Friday or Thursday, January 12. It was down, you couldn't even use it. And you had to like put it in, you know, for them to notify you because their servers are just so overloaded.

Matthaeus Huelse: 17:38

Yeah. I think that when we worry about plagiarism, which is obviously the first thought, I think what we need to really think about in the long term is what do we tell our students about the meaning of ownership of my my work my writings, pride in my own ideas and expressing them? Well, that it's not just about the concept of writing something for writing something, but for the purpose of learning, and for expressing my own ideas, it's, it's that same expression about chasing the learning rather than the task, and AI can help you finish the task. I can also help you learn, and that's what we need to put our focus on. But when it comes to what education is nervous about, and what what I think is something we need to keep in mind in is its bias. I think our our ideas of AI and what we think it does and what it doesn't do, there are so many inherent biases in AI that we need to be aware of, such as, I mean, the data that was practicing with is human generated, but the question is, which humans Yeah, right. Humans put that in there was represented. And that is something that we want to be extremely careful with when we talk with children, and students. I also heard that someone referred to someone referred to AI as a cheater,

Katie Ritter: 18:58

a cheater, cheater.

Matthaeus Huelse: 19:00

Because it is so good at recognizing patterns. There was an example where they used x rays for mammograms to figure out whether they I can recognize maybe cancer quicker than a human eye would. And it actually did really well they practiced on a set of data from one one hospital and it did great, it predicted did really well. And then as soon as you add the data from another hospital it It crashed. It didn't do so well. And we didn't understand why. It turns out that it wasn't looking at the actual mammogram like data that was in there. It kept looking at the little tag on the top left or top right corner of the X ray, recognizing the information from the tag from the what it said on there. And so we recognize that pattern. We thought it was doing great, but you'd recognize a completely different pattern we didn't even though Yeah, so it doesn't cheat but it is it can get distracted I would say

Katie Ritter: 19:54

oh my gosh, my I'm like really struggling to wrap my mind around.

Matthaeus Huelse: 19:59

Okay, Here's my favorite. Someone said that AI is a liar. And I would like to fight that opinion. I disagree. But I have to I have to in order to do this accurately, I need to know am I allowed to swear on this podcast?

Katie Ritter: 20:18

It's like the best s d are a word. Probably not more than that or won't get marked explicit. What are the H?

Justin Thomas: 20:25

Well, Tyler is the only one that's done that so far, and he didn't even say anything. But

Matthaeus Huelse: 20:29

well, what if it is the title of the book I'm referencing? Yeah, I'm sure there's going to be I'm sure there's going to be ugly. Okay. There's a book called on both by Frankfort, and he talks about the difference between someone who tells the truth and someone who tells a lie. And the thing that the two have in common is that they always are rooted in truth, the truth teller will always come from a point of truth, because that's what they want to convey. And the liar will always have to address the truth. The problem is AI is neither of them. AI is about AI doesn't understand what the truth is. It just gives you an answer to give you an answer. And it gives you answers that might not be correct. But our answers that you want to hear, it's worse than a liar. It doesn't know that it's lying. So when we are creating texts, we need to be aware of the fact that it doesn't quite understand what it's doing itself.

Justin Thomas: 21:18

I almost feel like this is another movie reference. But the the newest Blade Runner, isn't that have you seen the newest blade? Right? Isn't that what it Joy? Joy is literally feeding the because it's the pattern right that the main character Ryan Gosling's character has is that these memories they aren't actually his you find out the end. Spoiler alert. I should have said that beginning but I

Matthaeus Huelse: 21:38

gotta put that in there.

Justin Thomas: 21:40

Right. Yeah, exactly. But isn't that because it's not that she doesn't necessarily know that it's true or not, but she's just feeding what he wants to hear.

Matthaeus Huelse: 21:47

Yeah, that's the AI only deals with the data that it has. And it can only interpret it to some extent, okay, well,

Katie Ritter: 21:53

that's important for educators to know, less like scary AI. Bring it back to education, I think to like something for educators to get excited about Rachael Porter on our team shared that YouTube video of AJ Juliani, talking about the seven ways, educators can use AI, and a couple of ideas that he shared. I was just so excited about the potential for educators to us, in particular, you know, he shared how you can, you know, basically give it a lesson or a topic you're talking about and add it to incorporate, like UDL framework into it, right. And then it gives you these ideas. He talked about creating a cross curricular lesson, which I think will help education becomes so much more relevant because it's the only box where we do this siloed content for 45 minutes. And then five minutes later, you're doing this siloed content, and nothing is really connected well. So I don't know what the ability to you know, Matias, you're addressing the bias, which is absolutely, we need to think about the information and where it's coming from and who's who's representative, when we're building those neural networks, but I think the ability for educators to use this tool to address inequities in the classroom, and essentially duplicate yourself, and what you have the time for. I mean, differentiation is like this lofty goal that everyone who goes into education, like we believe we're going to do that, right. And then we see, educators getting overwhelmed with all of these other things they have to do. And like the reality of the education landscape today, the vast differences in like background and ability that students are coming to them in their classrooms today. I just like that made me so excited as like, you know, there's a lot of tools that say they're going to help educators and make their life easier. I feel like this that is the first example I've been seen that's like, fundamentally, or at least seen in a while, like fundamentally transforming supporting educators in a that a way that truly makes their lives better, in a way that will truly allow them to impact and support students.

Matthaeus Huelse: 24:13

I wholeheartedly agree, I think the biggest strength that AI offers us is that it can shave so much time out of our daily routine out of the things that we used to have to worry about. Like it can help us potentially with grading, which would be huge and can help us with lesson plan creation, idea creation really can help us with research. All of these things that I always worried about when I said to admin or anyone else like I don't know about giving our teachers the impression that we're putting something else on their plate. I don't know about worrying about this because they already have so little time. This is something that will potentially take give us so much time back.

Katie Ritter: 24:52

Yeah, yeah, I love it. And last thing I'll say before I break us to the sponsor break, I'll leave. I'll leave this one here, but I do just feel really strongly that for I don't mean to undercut or like not give credence to, like very real concerns, particularly with, you know, safety and Matias, you know, yeah, I think you said it really well. Like we don't even know the rules that we need to be putting in place just yet. So like, let me just kind of like flag that caveat, like, I understand that, I don't mean to totally dismiss that. But I think if our gut instinct is to just block, ignore block like we are only making school more irrelevant for these children who have very real access to these tools, or creating wider inequity gaps for our students who may only have access to these tools at school in comparison to kids, you know, if we're blocking them from it, then they get no access where we have other students. So anyway, I just I hope our schools are not going to make school more irrelevant by just like blocking and not addressing and not figuring it out. And Michael Roush our most quoted co Yes, on the podcast. He sent a tweet the other day, and he just kills me anyway, with his quick wit, but he said something like I'm hearing a lot of teachers talk about how much easier chat GBT is making their life but it's cheating for their students. But you know, but we put in like, give me this lesson frame. Give me an outline of how I might teach this. That's not cheating, right? Like that's helping support us. So but it's cheating if the students put something in and submit it to us. So I just you know, let's let's remember that. As we go into this quick sponsor break,

Brooke Conklin: 26:41

who is in your coach community. The edu coach network is a professional learning community designed specifically to help instructional coaches, connect, learn and grow together. With free and paid options available, there is a space for all coaches and the edu coach network. Join today and access coaching content that will help you impact teaching and learning in your school, go to www.edu. Coach network.com to join

Justin Thomas: 27:10

instructional coaches support teachers, students, administrators, and really everyone in the district. In fact, research shows instructional coaching is one of the most impactful forms of professional development that results in improved teacher instruction and student achievement. But who is supporting the coach Ford Edge provides multiple year long mentorship options recommended by the Google for Education certified coach program to help you gain the value and support you need as an instructional coach, visit Ford hyphen edge dotnet to start giving PD to the ultimate PD providers. Alright, welcome back, everyone. Justin Thomas here with Katie Ritter on the Restore recharge podcast. And we have Matias Holzer, the German expert in AI joining us here today and we're talking about Chet GPT. But before break kid, he talked about whatever was talking about how there's a lot of teachers that might be a little on fans a little nervous about some aspects. But there's so many amazing things that we've done as teachers that we should open up for our students. So we want to kind of put this in a positive light, right? So what are some ways that coaches can help teachers see the positives and the excitement around Chet GPT, or AI in general, and really kind of changed that narrative to be more of a positive outlook to actually be able to be used by everyone?

Matthaeus Huelse: 28:28

Absolutely. And I think I have a couple of good good tips. And the way that I approach it, I'll gladly share, I always try to open up by talking about the time saving benefits of it, because this truly should be the hook. When you when you realize how much time you'd be saving on lesson planning. I think that's how you wow them, you show them how it can create lesson plans, how it can help you with differentiation, and give them a lot of feedback on how to create ideas and use the chat JpT specifically for that, but you hook them I think with the ability to like summarize notes, to give it a little bit of input and give you feedback or summarization of it. And the last and final thing when it comes down to saving time, I always say it can teach you things rather than using it as a voice finder, which is also cool. But using as a writing coach, because if you use chat JpT it will ultimately do the work for you. But if you tell it specifically you are a writing coach, I am your student, I want you to read this input and give me specific instructions on how to be better, then it will not just changed the text for you, but rather will give you instructions like a teacher would which would be much more interactive, and you can then start asking it for the question. Well, why would I do it this way? Or explain to me what the benefit of this change would be? Yeah. So

Katie Ritter: 29:44

back to your point of like, you can learn things. Yes,

Matthaeus Huelse: 29:47

right. That would be a good way of using it. Okay. Okay. Then after that. address concerns. That's that's my next step because we jumped we talked about bias. We talked about how we need to be aware of what our students are exposed to, but I will also point out data privacy. The amount of access that the chat GPT has right now is one way might have some access to the internet. But what if it starts having access to more and we can ask it about specific people and their express it. Their expression on the Internet rules will change rules will change along the way, because we don't know them yet. And anybody that has the inclination of throwing student data with recognizable information into the Chatbot, please stop them. That sounds like a terrible idea. I don't know why you would throw student information that is identifiable into the chat. Yeah, that's about it. Yeah.

Katie Ritter: 30:40

Yeah. Good little asterix there. Okay, well, my tastes we end every episode with top three tips. But before I dive in to ask you about your top three tips, is there anything else that you'd want to say, for our educator listeners, or, you know, specifically with our instructional coach audience, or the main folks that tune into the pot here, just anything else on the topic? Before your top three tips? Don't give them away?

Matthaeus Huelse: 31:10

Absolutely. I really actually liked your point, Katie, earlier about. If we start blocking this tool that we are obviously comfortable using as adults, I also have a big issue with the fact that we would start locking students out in my personal opinion, the only people that are actively locked out of something by the district are usually the teachers and not the students. Because if the students communicate, they work, they will find

Katie Ritter: 31:34

a good point,

Matthaeus Huelse: 31:36

I don't think I ever had to help a student get access to Netflix, but I would have to work hard to help us teacher get Netflix on their computer. But it's not that we they don't get access, it's it's going to become a matter of who has the money to afford a smartphone with a data plan. And then it's not becoming about who has access anymore, but who has the money to get access? Yep. So I really, really agree with you on making sure that we access we give access to those tools that we would be using in the future to love it.

Katie Ritter: 32:06

Okay, well, then let's dive in. What would be your top three tips for utilizing chat GPT or, and or AI overall, in our schools.

Matthaeus Huelse: 32:20

So the first piece of advice I would give is that it is the wild wild west right now on the internet with artificial intelligence. There's lots of opportunity, very few rules, and humans adjust to rules in the tech space slower than the tech starts evolving. We need to keep in mind that if I would take you for example right now and not just throw you into the 1990s. Right, I'll throw you back in time from this podcast right now. Would you I look

Katie Ritter: 32:51

forward to staying home from school and watching Saved by the Bell, Dawson's Creek

Justin Thomas: 32:55

and prices right? Maybe that's

Katie Ritter: 33:01

a really great pictures of WhatsApp. Yes,

Justin Thomas: 33:03

yes. We are already there.

Katie Ritter: 33:07

Back in the 90s.

Matthaeus Huelse: 33:08

Would you be mad at me? Or would you say thank you, Matias for throwing back and throwing you back into the past and making you rich by realizing how much you can do with the internet that people don't even realize yet. Right? If I wouldn't be able to go back and take advantage of that. And I would want to take the most out of the internet, we are at a similar place and time again, there is a huge technological breakthrough. Don't give up on it, stay involved in it. And that's I think what we need to really keep in mind, it took forever for us to even get slowed slightly better about codifying issues about privacy and net neutrality on the internet. It's gonna take a long time for us to understand AI and what rules we need.

Katie Ritter: 33:43

Why am I excited Sorry to interrupt, make it quick. So you don't lose your train of thought on number three, but I'm excited for like what could be with education, particularly curriculum programs with AI, I just think like if it can learn then it could learn and tailor to that individual student for like true in the moment like support to push our students who need extra work, and they're ready for that next level to meet our students who aren't quite on grade level yet and give them you know, their level of appropriate tasks and to individualize feedback and support if they are getting it wrong. And to really understand like, what resource will help this student who's here in this moment so I you know, we see that in we I feel like we're seeing like how you said, we've been seeing like little introductory AI like the captcha and different things. And that's kind of like building AI. I feel like Google's practice set while not quite AI, I feel like that's kind of the mentality right of like, if I get the wrong answer in the moment, it will give me this resource related to this skill. And again, that's not quite AI making that happen. They're like pre vetted resources, but that feels very, like a preview of what could come so anyway, I like your advice to stay involved. Oh, I I think it's super exciting for for while education might be the most scared right now, I think it has huge, like future implications to like help our teachers in the moment with students.

Matthaeus Huelse: 35:11

Yeah, absolutely. I agree with that. And that actually leads me to the next point, I think. Because AI is going to start appearing everywhere in all of our production, like productivity suites, like both Google workspaces and Microsoft will, will have it included at some point in time, it's gonna go and appear and not drive throughs in our toasters, My piece of advice is, you don't don't trust it. I mean, okay, I'm not throwing us back into the Cold War, but I mean, trust, but verify is a very good rule for AI in my opinion, right. I mean, it's okay for it to generate information and it's good for us to use it and start off, but it's always a starting point. It should never be where we end it should be a starting point and never end.

Katie Ritter: 35:54

That's funny. Yeah, we're

Justin Thomas: 35:55

gonna 90s not like,

Matthaeus Huelse: 35:57

Oh, I've been all over the place. I've we've been witches we've been in the 90s have been on the road with we really

Katie Ritter: 36:01

have been on a story road today. Yeah, well, I think it's how you said we're gonna start seeing it everywhere. Brooke Conklin, frequent, frequent pod caster, who we just had on the episode, actually, last week, she threw in our Slack channel for the company about AI in education. AI is embedded in Canva. Yeah. So I haven't had a chance to check it out. But for those coaches, maybe, maybe chat GPT is a little over your head, like I'm feeling this very, but dive into Canva to get a little taste of it, to see how it can help you.

Matthaeus Huelse: 36:36

And there's lots of different ways that you can include it. And that's my actual last point, my last piece of advice is, there are so many different tools, there are so many different ways that you can include it right now. That include AI in your work, a good website that I would recommend is actually future tools.io. It has an app, you can click and start searching for different tools that are AI related or not AI related sometimes, but you can even look for picture imaging versus text imaging. But for those advanced users, check out the different tools such as tome to make presentations. Now that T O M, E, okay? And then the long term for AI and AI usage, what you would want to research and learn about is prompt engineering. So a big part of AI is we want to make sure we get the best possible answers. So a piece of advice and prompt engineering might be tell the AI who it is first, tell it who you are, and tell it what its purpose is and then give it the information to work on. And these are the kinds of rules that we are learning as we're going through this.

Katie Ritter: 37:40

Yeah. So give me give me an example of what you mean, I think I know what you mean. But just to flesh it out. Give me an example of what you say like tell it who they are, or what it is.

Matthaeus Huelse: 37:48

Okay, in the classroom environment, what I would suggest using it for is, as a voice finder, for example, I will take a whole bunch of my writing different kinds of writing pieces, articles, whatever I wrote. And then I'm going to tell it, you will be a writing coach, you're going to look over my assignments that I've written my articles, and you're going to write a quick summary, what my writing style would be described as, so then it'll read the text, it'll give you a quick summary, I told you, this is your job, this is what you do, please give me a response. Once you have that response, I use that voice find out that little description you can take that go into the AI playground, which is slightly different from Chad GPT, but gives you more opportunities. And you can adjust it and teach the AI to behave and write like you like your writing style based off of that, because it'll start saying that you're right in this particular sentence, very matter of fact very something rather. And then it'll start recreating your sense of writing

Katie Ritter: 38:47

Wowzers Okay, throw that in the cauldron too. And I think on that my head has officially exploded. And that might be all we can take about this AI here today. Justin.

Justin Thomas: 38:59

Okay. Before things get too out of control here, but Mateus thank you for joining us here today. Yeah.

Katie Ritter: 39:08

Sorry, you're blowing my mind. I'm getting very educated.

Justin Thomas: 39:11

Yeah, so plenty of amazing ideas and some, some really good tips there at the end too. And kind of how you can really use AI and chat GPT and go forward if you're a coach, helping teachers kind of understand what it is I'll use a little bit more and

Katie Ritter: 39:27

I don't think even my Botox could keep my eyebrows down

Justin Thomas: 39:33

and also using it to you know, easier your own life as well. Well, we have next next time out here on the on the podcast, we I mean, we got several good episodes coming up. We're gonna have one we're rolling to TCA. We're gonna have a discussion and kind of on some of the sessions. We've learned from that. We also have another one coming up as well, in terms of kind of how you can as a coach, make sure that you are able to kind of have your time and show up is that you're doing some amazing things with the time while also not like pushing boundaries with teachers, especially if you're at some of those smaller districts where you know, you're seeing the same people like every single day, but we got some good episodes all coming up so make sure that you're still listening to the restore recharge. Yes,

Katie Ritter: 40:14

absolutely. And I believe when this episode airs, Justin that our coaches camp you heard an ad for the early bird pricing is ending in March so be sure to register for coaches camp if you haven't yet. And with that, I think that wraps this up here so be sure to subscribe the restart recharge wherever you listen to podcast and follow us on all the social media at our our coach cast,

Justin Thomas: 40:40

feel free to reach out to us to if there's any topics that you want us to discuss. We are more than happy to listen in and see if there's another way to do

Katie Ritter: 40:49

it. So press the restart button. The chats up here. I know I tried to think of a joke and I couldn't wait with the closing line.

Justin Thomas: 40:59

And then I jumped in. Alright, so take to the restart button recharging coaching batteries and leave feeling equipped and inspired to coach feel asleep with a restart recharge podcast,

Katie Ritter: 41:12

a tech coach collective