[audio logo]

Speaker 3: OTAN, Outreach and Technical Assistance Network.

Neda Anasseri: All right. Hi, everybody. And welcome. Thank you so much for joining us today. I'm very, very excited to open up this webinar and introduce our friend Rachel Riggs. She'll introduce herself a little bit more when I hand it over to her.

But what we wanted to do for you today is partner with our wonderful friends at World Education and the EdTech Center to bring to you today's webinar. This is a two-part webinar. This is part one of CampGPT.

And you'll be joining us for the second part. If you haven't already registered for that, we will definitely remind you towards the end of this webinar how to register. We thank so much World Education and the EdTech Center for doing this webinar series for us. We really appreciate it. And I'm going to-- enough from me. You're really here to see Ms. Rachel Riggs. I'm going to hand it over to Rachel. Thank you, Rachel.

Rachel Riggs: Thank you, Neda. We are always happy to work with OTAN too and the amazing educators in California, so thanks for having us. Before I even introduce myself or explain what CampGPT is, I want to hear from you guys.

So I have a little poll. I just put the link in the chat. I also love the introductions coming in. So hi, Jennifer. Hi, Virginia. You guys can continue to introduce yourselves. But we're going to get into Mentimeter. So if you go to menti.com, you can enter this code, 73033019, or you can just go ahead and open the chat box and click on the link that's in the chat.

And I have a few questions for you. I'm going to pop up the QR code. For anyone who likes to use their phone, you can scan the QR code. See how I'm giving you so many points of entry here. So if you prefer to use your phone, scan the QR code. We have a few questions about your relationship with robots, nothing too personal though.

All right. I'm trusting we've got people coming in. I'm seeing some thumbs up, which lets me know people are getting in OK. Thank you, Neda, for putting the code in the chat. All right. And you can continue to join. So I'm going to go to our first question. But if you're not in yet, that's OK. You can still click the link in the chat.

First question is rank your knowledge-- I'll put the QR code up here. Rank your knowledge of generative artificial intelligence, or GenAI. We've done this now in a few different states. You know I've been thinking it would be funny to compare state against state and make this into a competition. Who's learning about generative fastest?

I can see our California friends are well far along. We have a lot of people using it. You can choose multiple too. So if you're using it and you know strategies, you can indicate both of those. So let's see. Lots of I have used it. OK, cool.

Good to know. I can't wait to hear from you guys. I have heard of it. I can define it. That's awesome. A little bit less of that. That's OK. We will define it together today. I will share strategies with you today. And I will also give you guys the opportunity to share your strategies as well.

And this is great to see. Our session today is designed for all levels. So you are all welcome. We are all going to build our knowledge together because today will be very collaborative. And we will have some hands-on work after today's session so that you can further develop your skills.

So thank you all for joining. Thanks for letting me know what you can do, what you know about generative AI. Now I want to know, if you could ask a robot-- I'm going to get really imaginative here-- to do anything this week to make your job easier, what would it be?

So really think. What do you really wish the robots could do for you? I Don't feel like they can do the things right now that I really wish they could do. My monthly budget or slapping my hand every time I go to Target. That would be nice. Yeah, clean my house well. That would be nice. Dust my house. Oh, my job. I love that.

Every time we do this, by the way, it is usually majority not work-related. It is usually majority like clean my house, do my other tasks for me. So that's interesting. Cook for me. But we do have some here related to education, build a module in Canvas.

Speaker 4: That's something you can do.

Rachel Riggs: Make worksheets. Yeah, OK. Ask for a step by step lesson plan. Oh, you guys, love it. We're going to get there today. Update all EL Civics materials. I want to know who put that. I want to know who put that.

OK, create an accessibility-compliant presentation. Oh, yes. Yes. Yes on that one. Somebody talked to big tech about making that easier on us. I recently spent such a long time. I had created something in Canva. And I put it in Adobe to check the accessibility and then to retag it. And like, wow, you guys, if you want anything to be accessible, don't make it in Canva. That's my only advice.

Create a schedule for a two-day conference with 60 to 80 session proposals for another session. I think new one from design a logo. Awesome. Have a couple of difficult conversations. Interesting. OK, cool. Thanks for sharing.

So these are all the things we would love for a robot to do for us. I can't come through with all of those. But I think some of them we might get into with our use of generative AI. What do you want to learn in CampGPT? We have a lot of different levels-- a lot of levels of experience. And so I'm sure you all have different goals. Tell me what they are.

What do you want to learn? I'm catching up in the chat. Good to see you. I see some familiar names. Yay. Clean my pet's waste. Someone said in the chat. I agree with that. What do you want to learn? Tricks and new ideas, best practices, print clarity. Whoever put print clarity, get into the chat and tell me what you mean by that because I'm not sure.

Keep it coming. What do you want to learn in CampGPT. Now is your chance to tell me. What and how to teach my students about AI. OK, you have to come to the next session. We're going to talk more about that. But today, as we're talking about teaching with AI, I think it will give you some ideas too.

All right. Learn ways to make EL civics assessments updated. Learn how to translate course content into various literacy levels. Learn how to use it better. Optimize my time. Train others about the basic AI concepts. OK, cool. We can help with that. I can certainly also share some materials with you if you want.

Create effective parameters. Good resources for teaching. Use its power for good, not evil. Yeah. Ensure chat isn't hallucinating or biased. We'll talk about that for sure. Prompting. OK. And hear what others are working on. And honestly, I'm going to be tapping into you guys a lot. I want to hear what you guys are doing, what your strategies are.

So this is all good to know. And like I said, it'll be very collaborative. I'm going to address some of those things. But more importantly, I hope to facilitate a space where we can all be sharing our knowledge and skills. I'm bringing up my slides now. My computer's going a little slow. I hope you guys can hear me OK. I figured out to use the clips.

Neda Anasseri: You are coming in clear, Rachel. We're all--

Rachel Riggs: OK, good. Oh, prompt clarity. I see, Babs. Thanks for clarifying that one. Excellent. We'll work a little bit. I'll show you guys a prompt framework that I like today.

So welcome to CampGPT. I'm so glad to have all of you. And I am Rachel Riggs. It's nice to meet you. I'm a technical advisor for World Education. We're a nonprofit with a mission toward equity in education around the world.

Here in the US, we are specifically working with adult literacy learners and English language learners. My role spans across a few different initiatives, first, the EdTech Center where I do a lot of teacher training and capacity building around EdTech integration and digital skill building.

The CrowdED Learning initiative is around open educational resources. So I work with teachers on designing and evaluating those. And then our AI for learning and work initiative, which I lead. And it spans teacher training as well as working with other stakeholders on AI literacy and the transformation of learning with AI. Buzzword and buzzword and more buzzwords.

So in CampGPT, our goal is to bring in some principles around the ethical use of AI, some practical strategies for using generative AI, a lot of sharing and collaborating since this is a new space for all of us, and also thinking about how we're going to build the skills that learners need to navigate this new AI-infused landscape.

We have three sessions of CampGPT. Today is the first one. We'll do a little camp orientation. And we'll talk about teaching with AI. February 16 is our second session. And I'm hoping some of you, after you do some asynchronous work, will be willing to do a little show and tell on February 16. And then we're going to talk about teaching for AI. And that's around what learners need.

Finally, we'll have a showcase at TDLS on March 2. I'm also hoping to have some copresenters for that. So if you're interested, let me know. And that will be to share the work that we've done together here.

So three sessions. Getting started today, we're talking about a teacher's role in the age of AI. We're going to talk about some AI ethics, which we present through our fun and kitschy camp rules. We'll go over some prompting strategies. And then I'm going to give you guys an assignment to do between now and our next session.

Let's start with just some basic definitions. What is generative AI? Which brings up, I think, a lot of other questions like, what the heck even is AI? Which even computer scientists don't agree on. So I like this graphic. I do think it helps.

We can think of AI as the big field that is very theoretical and philosophical, even includes aspects of math and statistics and so on and so forth. It's a very huge field that spans many different disciplines.

Then we can think within that, there's machine learning. And that is just the general ability of a computer to learn from data. Then we can think about deep learning. And when we talk about deep learning, we're talking about a more complex, what they call a neural network. So how computer scientists have taken the brain and the synapses in the brain to make machine learning even more effective through this concept of deep learning.

Within deep learning, we have generative AI. So this is what we've been hearing so much about. It has been around for a while. But it's gotten really good. And it's been released for public use. And therefore, there's a lot of hype right now and a lot of interest rightfully so in the capabilities of generative AI.

Generative AI is unique because it can generate new and original data in the form of a text, image, audio, video, et cetera. So it is taking the data that we put into it. It's using algorithms, deep learning algorithms, and various parameters to smush and learn and develop something that is new and original but based still off of the data that's been put in.

So like I said, a lot of times when we say, What's generative AI? it brings up a lot of other questions. These are some acronyms you may see. An AI is artificial intelligence. ML, machine learning. DL, deep learning. GenAI, generative AI.

I wonder if anybody else can answer these. What's NLP, LLM, GPT? I'm sure we have friends. Patrick said homework. Yeah, homework. Good, Anthony. Got NLP, natural language processing. So processing and outputting language. That sounds very natural and is almost indistinguishable from a human. It's indistinguishable, I would say.

LLM. Yeah, 100%. LLM is a large language model. That's what tools like ChatGPT are built on. And GPT is generative pretrained. And I'm not even going to define this because I don't even think it's applicable to our work as educators.

But basically, like I said, generative artificial intelligence can generate new data. And so when we think about what we see in the different applications, ChatGPT, which I think we all know, is a chatbot that generates text and now can generate images and can output things like titles, outlines, stories, poems, and images.

Midjourney is a text-to-image generator. So here you put in a text prompt. And it generates an image for you. Synesthesia is a video generator that manipulates an avatar based on a script that you input and gives it a realistic voiceover to create a video.

So these are just some examples of different things that you input with a generative AI-based tool and the things that it can output, which are growing and expanding every day. We're going to talk more about chatbots.

We're going to talk a lot about large language models and chatbots throughout CampGPT because that's where a lot of teachers and especially adult literacy educators are finding a lot of potential and use. But I will say in your assignments, you are, of course, welcome to experiment with generating other types of media.

So what is a teacher's role in the age of AI? According to the European Union and their Digital Education Hub-- OK, we have questions. OK, I don't want to miss questions. I'm writing a course on-- would it be ethical to put a YouTube clip and ask AI to write literary analysis and/or assessment prompts?

I'm going to save the ethics questions because I'm going to get to that a little bit. Yes, you will have access to the PowerPoint. Thanks for asking. We will share this out. So three steps for educators. Number one, teach for AI. Number two, teach about AI. And then teach with AI.

So for AI is for user perspective. Knowing that learners are users of AI-enabled tools, what do they need to know? About AI, number two, developer perspective. So thinking about learners' potential to move into a career where they would develop AI models, what do they need to know?

And then three is teaching with AI. That's about us as educators. How do we appropriately apply AI-based tools in the classroom. So when we think about teaching for AI-- and I want you guys to be thinking critically for a minute here because I'm going to ask you a question after I talk about these three things.

When we talk about teaching for AI, these are examples of what that might look like. It might look like integrating data privacy concepts into a social studies lesson on human rights. Teaching for AI. Thinking about learners as AI users and what data privacy they need to be aware of or data privacy issues.

Integrating algorithmic literacy into a language lesson in which learners are describing the steps of a process. This gets to the computational thinking that we all need in order to really effectively use technology. So those are two examples of teaching for AI.

Teaching about AI is a little more explicit and maybe what you might consider hard skills. So when you're teaching about AI, you might be having learners create an application to learn about classification and data sets. So you can see that's very AI-specific or there's-- and these in parentheses, these are like sources of these actual lessons that do exist.

From aiEDU, they have a whole set of lessons called AI Olympics where learners are actually training an AI model. So again, just a few examples of what teaching about AI would look like. It's more from that developer perspective.

And then there's teaching with AI. So here, it's about a teacher's relationship with AI and how we apply AI-based tools in the classroom. Here, we might be using Bard, or ChatGPT, or Claude, or any kind of chatbot to create or update a syllabus, or we could be incorporating the use of an intelligent tutoring system to personalize instruction. So it's about how we are using AI to enhance our own craft as teachers.

So I want you guys to think this is what the European Union has as the hierarchy here where teaching for AI is prominent and most important, and then teaching about AI and with AI branch off of that. Do you agree with it? How could or should it change according to the needs of adult learners? You can ignore any ELA classes in any class-- math, ESL, whatever you teach, civics.

So what do you think? What's important for our adult learners and for us? And you can say, I agree. That looks good.

Speaker 5: Are you wanting us to change the hierarchy?

Rachel Riggs: Sure. Yeah.

Speaker 5: So what I do is I teach with AI. That's the top of my hierarchy. I haven't really considered the others too closely.

Rachel Riggs: That's great, Babs. That's good.

Speaker 5: OK.

Rachel Riggs: Yeah, thank you. Any others? What do you think?

Speaker 4: And I agree with her 100%. I'm using it. My students are all in prison, so it's-- maybe I shouldn't answer this. So--

Rachel Riggs: No, you should. That's valuable. So I think that's what I've heard most often when asked this question. It's most important that I know how to use it. How can I bring it into the classroom if I don't? So good. So it sounds like we collectively might move teaching with AI up a little bit further, then teaching for and about AI. What are your thoughts on that?

Speaker 4: As I'm discussing this with colleagues who teach, as I say, on the streets in the real world, I think it's forcing educators to ask better questions. So we've gone from where we-- way back in the dark ages in the 1980s when I started, you asked very rote knowledge based questions. And now we're having to-- I ask higher-order thinking questions. And I think the AI ability is forcing us to ask better questions.

Rachel Riggs: That's good, Patrick. So I'm hearing what you're saying. And I'm wondering how-- so if we're talking about engaging students more in critical thinking in order to navigate and use AI well, which would that fall into teaching or teaching about?

Speaker 4: You're asking me probably about.

Rachel Riggs: Wow. OK, interesting.

Speaker 4: I think the four is how to ask the right question. And I think we're already doing that with Google. It matters how you ask the question or if you have the vocabulary to ask the question. My metaphor for that is when you're having to go into the hardware store and you need the widget presents.

Well, you're never going to find the widget presents. You have to use the right name of whatever it is you're looking for. And you might not know it.

Rachel Riggs: I like that a lot. Yes, having the right language.

Speaker 6: Hi. Excuse me. This is Melinda Holton. I'm support. Kunyi, you need to mute. And everybody else that has just come in, please mute your microphone so that we can have a good session here without any interruption. Thank you.

Rachel Riggs: OK, all good. All right. Thanks. So what I've talked to adult educators in the past about and keeping in mind that this hierarchy was created for primary and secondary educators, so more like younger learners, I've definitely heard that that resonates about teach with AI. That's what we need first, is what I usually hear.

And then I also hear that teaching for AI, teaching those skills that learners need from a user perspective is something that is really important to adult educators. Teaching about AI, they feel it's almost like a dotted line. That's how I've heard it described if you think about this concept map. And there's a dotted line where teaching about AI is optional, depending on the learner's goals.

We know that not all adult learners are trying to be developers or software developers or programmers or computer scientists. And so that should be-- developing that developer perspective maybe is something that is optional and depends on their ultimate goals.

OK, great. Well, good discussion. Thanks for sharing your thoughts. And I want to say too that teaching with AI is our focus of today. And then teaching for AI will be our focus of our next session. And I'm not going to do teaching about AI because I won't be good at that because I don't have the developer resources. I can share resources for those of you who want to teach about AI.

But we're going to focus on, how do we as teachers develop the skills we need to teach with AI? How do we pass on some of those digital literacy and critical thinking skills that learners need to use AI-based tools? And then I will share resources if you are interested in teaching more about classification, data sets, training AI models, algorithms, and some of those more technical aspects.

So teaching with AI. These are some of our key practices. These are our camp rules. And they align to-- OK, I'm catching up in the chat. It's like Excel. Yeah, Patrick, I like that comparison. An AI is a lot of math too. So it's like, OK, I know how this works, so I can understand it and use it better. But I may not need to know all the technical-- all the formulas to use in Excel, right?

OK, great. So teaching with AI. Some of our key practices here. We have our camp rules, which spell out GEAR, so your GEAR for camp. And it's an acronym as well for our rule. So G is goes before tools. And it aligns to this AI concept and the EdTech integration best practice of having a strategy and a purpose before you adopt technologies.

Explore and have fun is really getting at your confidence and the experimentation that goes into this. Avoiding bugs is about knowing the pitfalls and avoiding them. And then remembering to buddy up hints at the human-centered aspects of adopting in your teaching practice.

So don't forget your GEAR. Let's break these down a little bit further. And let's look at them with pretty emojis because that will help us remember them. Don't forget your GEAR. So number one rule, goals before tools. Goals before tools gets at this idea-- thanks, Neda. I'm glad you like it. Gets at this idea of aligning.

In the Office of EdTech report, they talk about aligning AI models to a shared vision for education. I like to tweak that a little bit and say, we need to align our AI implementation to a shared vision for education. So it's about taking what is our vision and our goals, and then only using AI when it makes sense to further that vision.

So what is our vision. You might envision that high-quality education for adult learners as your vision, maybe more contextualization so that learning is relevant to adult learners. Maybe your vision is that education is free for all.

What is your vision for adult education? Can you guys share? Thank you, Anthony, for sharing the report. Can you guys share in the chat, what is your vision for adult education? Let's start there before we start thinking about-- OK, thanks, Maria. Accessibility. Good. So you should go into today and go into all AI use with that vision in mind. How will AI actually make learning more accessible for my learners? And it can't.

Free or low cost. Accessible and equitable. Patrick says lifelong learning. I think getting back to those cognitive skills you were talking about, Patrick, so good. You have to think about, how will AI help us work for that vision?

Student-centered, useful, personalized, customized. Yes. We know AI is good at that. Giving us more options. Developing materials in other languages at different levels. Help with the work. Help with the work. OK, I like that, Aaron.

Some people refer to that as augmenting intelligence or augmenting our work rather than thinking about it as something that will replace what we're doing. It's going to support what we do in our learning journey and in our teaching journey. Great.

So here's an example of goals before tools. What we know is that in adult education, we want to provide resources that are highly contextualized, that spurs motivation and persistence on behalf of adult learners.

So an example of putting our goals before tools would be me thinking, you know what, I want to develop a lesson plan on writing a compare and contrast essay that integrates environmental conservation, particularly reduce, reuse, recycle.

So I'm thinking about the academic skills, compare and contrast and writing. And then I'm thinking about a specific context like green skills or the environment. And I'm wanting to have a lesson plan that will help me with that.

So if my goal is contextualization, what I know through my own experimentation is that I get nice creative ideas from ChatGPT. So I would choose ChatGPT as my tool for developing this lesson plan. And I've aligned that with my goals.

Now, if I have a different goal, maybe my goal is to use generative AI tools to support learners in their writing, what I know is that all of my learners have access to Google. And I know that Google's tool, Bard-- that's their chatbot-- is really good at pulling concise, less creative but more factual information because it's connected to the Google search engine.

So in that case, if I'm looking at differentiation and I know these different learner factors, I would choose the tool Bard for my students to have that writing support. So this is an example of how I might use it with learners.

If they need help developing their outline, if they need help during prewriting, I would encourage them to log into their Google account, use Bard, and ask it to help them develop an outline for their essay. So two examples of two different goals and choosing two different tools that align with those goals.

But how did I come to those decisions of choosing or choosing Bard? Through experimentation and learning. So part of this process is being willing to explore, experiment, and have fun. This gets at what we call digital resilience.

So digital resilience is the idea-- I think most of you probably have heard of digital skills, digital literacy. That's getting at what we're able to do with technology. Digital resilience is about how we're able to keep up with technology in a sense.

So it's about our awareness, skills, agility, and confidence to be empowered users of new technologies and adapt to changing digital skill demands, which we know that those demands on teachers especially have been really intense since about March 2020.

So digital resilience is part of what we're practicing here together, learning new things and being willing to explore, experiment, and make mistakes in order to develop new digital skills and adopt new technologies.

Yeah, Jennifer, I'm sorry. I know my little-- oh, there we go. I got it. I was about to say I'm having trouble getting these links but I'm not. And I'm going to share the slides too and like each-- if there's any references or resources in the slides, then you'll be able to click these links and explore more.

So digital resilience improves our capacity to problem solve and upskill and navigate digital transformations, which generative AI has been a massive disrupter and transformation, and be active participants. So we as teachers want to develop digital resilience. And of course, we want to support learners in developing it.

So part of that is experimenting. I have a chart here. Actually, this one might-- let me make sure this is the right link for you guys. This is bringing you to another Padlet. I forgot to transfer this over to our Padlet. But it doesn't really matter. You'll find in the chart. And I can put it in our Padlet too.

But this is a chart of different generative AI-enabled tools. So we have some of the popular tools from big tech like ChatGPT, Bart, Claude. And then we also-- generative AI has been built into some of the popular EdTech tools that you may already use like Canva, Khan Academy, Quizlet.

So this chart is meant to just give you some ideas. If you haven't started experimenting yet, it'll give you some ideas of tools that you can experiment with. But the point here is to feel confident and comfortable just experimenting and checking out what these different tools can do.

And also acknowledging that all of this is part of a journey. So for example, when I enter a prompt, like, I teach an ELA class in which they're blah, blah, blah, blah, blah, blah, blah. This is prompt. Asking a chatbot for different words and a word family.

ChatGPT, the free version, will give me one response. Claude will give me another response. And Bard will give me another response. So there's a huge amount of variety in these generative AI tools and in the output.

If you're working with image generators, for example, you will struggle immensely to get the same looking person out of an image generator. It's hard, unless you're using a really specialized tool.

So just know that this is also part of the nature of these generative AI tools, especially the ones that haven't been fine-tuned and specialized for a specific purpose, is that you will get a huge amount of variation in the responses. So this experimentation is essential to learning how the tools work and the nuances between them.

And then we also want to avoid bugs. And so some of the drawbacks of generative AI include data privacy. And so we have to think about what you're sharing when you're using chatbots or image generators or any of these tools. What are you comfortable sharing? What do you know about who you're sharing it with? And what do you know about how it's being used?

I think that's pretty self-explanatory. But you just want to be careful. Usually, it's hard to know who you're sharing with. And it's hard to know how it's being used. So most of it will fall into the category of, what are you sharing?

I will go into ChatGPT, for example. And if a prompt that I'm giving it or I'm working with it and I have specific names, I'll change the names from Rachel Riggs to John Paul. And that way, I just feel a little bit more comfortable about the information that I'm sharing, the personal information.

Another aspect of avoiding bugs is hallucinations or better known as just straight-up inaccuracy. So what we know is that these tools will generate things that are not-- thanks, Anthony. You're the best. I'm sorry.

So what we know is that the tools aren't always accurate in their response. This is an example of what an image generator gave me. And I think my prompt was something like facts. It was one word. And it was very ironic. It was like-- I forget exactly. But this is what it produced.

What is this? This is something that maybe could be like a science anatomy diagram. But when you look at it closely, it's complete nonsense. There's not even a single word in here that makes sense.

So this is a visual representation of also what chatbots are doing when they give you language. They're meshing and mashing things together that sound good but ultimately are not factual and accurate.

So you want to think about what information does the tool have access to. So Bing Chat, for example, if you have a Microsoft account and you're in Microsoft Bing, which is their search engine, and you're using that chatbot, that typically has access to some of the personal information connected to your account so that it can give you relevant results.

So for example, if you're talking to Bing Chat and you ask for a pizza place, it'll probably find one maybe that's near you or at least in your city because it has access to that information. So you want to think about, What information does it have access to? to think about then, how could it be accurate?

And then also, how will you verify the output? So somebody mentioned earlier something about leveling texts in Lexile levels. What we have found through some experimentation is certain tools are better than that-- better at that than others.

But it's not always totally clear which tool is good at it. So for example, you might go into Bard and say, hi, do you know what Lexile levels are? And it might say, yeah, I'm really good at that. But then when you ask for a text at a certain Lexile level if you actually did the analysis of that text, it wouldn't be at that level.

So it's really important to think about how you're going to verify the output and what is an appropriate use because in some cases, it doesn't matter. Maybe it really doesn't matter the Lexile level. You just want it a little more in more simple language.

And that's fine. Then you probably don't even need to verify, except through your own eyes and reading. So thinking about how-- what information has access to and how you verify. Thinking about bias and stereotypes. Again, this is something that comes through really clearly in an image. But it also applies to chatbots in the language that chatbots produce.

So this is an example of-- let me put the link in the chat. I'm going to be a good host here this time. This is a great study that was done. There are a lot of studies similar to this, so if you're interested, I could share those as well. But basically, it's looking at Midjourney, which is a tool I mentioned earlier.

And when you put in the prompt an Indian person, it's coming up with some Native Americans in headdress. It also came up with-- I mean, if you look at the other gray images here, I think there might be one female in there. So you can see visually how the bias and stereotypes come through.

And when you're working with a chatbot, those will also come through. It just might not be as transparent to you. So thinking about how you can read between the lines and see where there may be bias, especially if we're talking about serving diverse learners and wanting the material that we use with them to be representative of their diverse backgrounds and goals, et cetera.

Finally, remembering to buddy up, which is my favorite thing. This is all about human-centered strategies. So thinking of yourself as, first of all, the most important human in the picture and how can you work with a chatbot, prioritizing your own expertise, and making sure that you get what you want out of it?

Human verification is very important. So if you're working in IT and occupational skills, who needs to verify that indeed what the chatbot has produced as for a practice exercise or something is the right terminology for that occupation? So also thinking about, again, human verification. Who are you going to run those by?

And then sharing transparent is another important human-centered strategy when we're using generative AI or any kind of AI, being transparent about how you're using it with learners and how it affects what they are interacting with.

So I like these for-- I mean, especially for the bias. These are two resources for checking for biased content. So if you're thinking about using generative AI to develop lesson plans, quizzes, or to develop stories or any of those kinds of materials that you bring into the classroom, these could be really helpful if for nothing else, then to give you a framework from which you can be more effectively verifying the materials that you use or evaluating the materials that you use.

Another aspect of being open is everything that we're doing here in CampGPT together. So we are going to be sharing a lot with each other, the work that we do with these chatbots. And that is part of us being transparent with one another and developing a community around our use of AI.

So we will have-- this is a worksheet that I'm going to share at the end of today's session for you to do your experimentation and then give it some structure and share it back with the rest of the group.

And then we also have the Open Prompt Book. This is something that we developed in a past CampGPT. It has prompts from actual adult educators and links that you can use. If you want a starting point, you don't know what you want to use these tools for but you want to see what they can do, this is a great place for you to start.

You can click on ChatGPT or click on Bard, and then you can just view someone else's conversation with these chatbots before you dive in yourself. So there are open prompts that you can look at and concrete examples for you of how other adult educators are using these tools.

Another part of sharing transparently goes beyond just our collaboration as educators. And it goes into this concept of what we called teaching for AI. So if we are developing the skills that we need to navigate AI, we are getting more acquainted with AI, we are using it to automate some of our tasks and really leverage its benefits and its potential, that's something that we should also think about passing along to our learners.

So what is the potential that it holds for them? And how could they be using it? And then how can we help them develop some of the same skills that we're developing in terms of data literacy and verifying the output and that information literacy piece? Like I said, we're going to get more into this in the next session. But it is an important aspect of this sharing.

So we're going to reflect on these rules. I know that was a lot. And maybe you were thinking to yourself, what the heck? I came here to learn how to really use this stuff. But we want to set this foundation. And I'll be working ethically and working together with a common understanding.

So that's what these rules are about. But what I really want to emphasize about the rules is that you're probably already applying them in other ways. So I want us to jump into this Padlet together. And I want to hear from you, what are the ways that you're already implementing these Camp Rules in your use of educational technology?

When is a time that you have chosen one tech tool over another because it was more aligned with your goals? When is a time you explored and experimented and had fun with technology? When was a time you had to create a workaround because you knew a certain technology had a bug or a pitfall?

When was the time that you relied on other educators, remembering to buddy up and the human-centered and collaborative and transparent aspect? When have you put that into practice? So you guys add your answers to the Padlet?

I'm going to take a break from talking your ears off and see what we've got. You can add your answers. And you can also like other people's answers, which is nice.

So if you're new to Padlet, you click down in the right where the plus sign is. And you can type here. Now it's finding that at home. So you can type here. Type your response to any of these questions.

So let's say explore and have fun. I'm going to say, Zoom, during the pandemic, I opened empty Zoom rooms to test out features. And then I can select where I want this to go. So I'm to have that go to explore and have fun. And I'm going to publish.

Yes, I love this. I love what somebody put here about exploring and have fun, having a teacher account and a student account. Definitely true with some of the EdTech tools. I would say like with some of the more open-ended tools like ChatGPT, you're going to learn probably everything you need to learn just from the teacher account because it doesn't have really a student side.

But I with Camigo, it has a toggle. So you can actually have the student experience if you're logged in to a teacher account. And we see how people are already experimenting with ChatGPT-- that's awesome-- to create daily readings. Love that.

Barely learning about all the options out there. Good. Start with your goals. And work from there. You have to try the others. I wanted to image for a lesson. OK, there you go. Avoiding bugs. I wanted an image for a lesson. But the face came out wonky.

Spent four hours trying to convert a QTI file for Canvas. Correct. So frustrating. That's another bug that I think we don't talk about enough that we should, which is thinking really and having a very evaluative approach to these tools.

So we want the promise to be that it's going to save us time. And it's important for us to go in and really measure that and think critically about that, Is it actually saving us time? is a really important question to ask ourselves.

Within three hours of Neon introducing DLAC to ChatGPT, I lost an entire evening. Yeah, that goes into-- I'm going to file that and explore and have fun. That's part of that exploration. OK, write assessment. So that's an example of that goal before tool. So if your goal is to develop assessments, we have to find what tool best aligns with that goal.

Chatting with other teachers, absolutely. Everything we're doing here as part of this buddying up, talking to other teachers, hearing how you're using it, what are your tips. Super important. MagicSchool.ai is like a big one. Teachers love this tool. It's a little more fine-tuned for teaching purposes, obviously, MagicSchool.

So that's definitely one to check out. It's not on my chart yet. It's one of those ones that's in my queue to add to my chart. So I'm glad somebody mentioned it. People love MagicSchool. Excellent. Goes before tools.

Chose ChatGPT because our district AI group advised that ChatGPT is more robust and a good go-to. Yes, ChatGPT has been proven in terms of its language capability to be a lot more capable than other chatbots. But there are tons out there.

I've found that sharpening prompting skills to target exactly and create what I want. So just experimentation with different prompting skills. And that is a way to avoid bugs too. If we think about image generation and how just the addition of an adjective, like if we looked at that an Indian person example, an Indian woman, adding just those adjectives, that's part of the human-centered too.

I am using my expertise, my goals. And I'm telling it more directly what to do. I'm leaving less up to the AI for it to decide. That's a great strategy. So let's talk about some prompting strategies then. I know we only have 9 minutes. That's OK. We can do this in 9 minutes. I'm not scared.

So let's talk about some strategies. So this is an example of the ChatGPT interface. It's actually a little bit of an older interface. I can't possibly keep up, you guys. But some of the capabilities that you have when you're in the chatbot, you'll have a menu usually on the left that's usually pretty much the same for all the tools. That's all of the conversations I've had with ChatGPT on the left.

What you're seeing in the center here against the white background is my prompt. And then against the grayish background on the bottom is what we call the output. So it's what ChatGPT gave me. So what I can do is I can go back and edit my original prompt. I can return to other conversations I've had. And I can give feedback, which is very important. I can give feedback to the chatbot. I like it. I don't like it. And I can copy the result that it gave me.

In Bard-- this is another chatbot-- I can Google the response that it gave me. So I can go see more sources by conducting a Google search based on the output. I can share. And then I can even tweak the tone and some other aspects of the output so that it'll give me something new.

I don't have to do another prompt. I can just slide some little customizations there. And it'll give me something new. I can also just keep the dialogue going. And so these different features feed into some of the strategies that we want to use when we're using these chatbots.

Effective prompting and detailed dialogue. So effective prompts, one of the biggest things that you'll hear in terms of-- and I do not use the term prompt engineering. I'm antiprompt engineering. But that is a term that you'll hear. I'm not anti, you guys. Somebody was just like, what?

It is a term that you'll hear. But I feel like it's too like-- whatever. We're just writing prompts to engineering for me. So using a prompt framework is a great place to start because then you're giving something really robust for the chatbot to work with.

This is when I like. It's called RACEF. There are tons of them out there though. The biggest thing here is just thinking about what details you can include to get something better. So RACEF has role. You're telling the chatbot what it is. You are an adult educator.