001Well good afternoon everybody. Welcome to our school board special meeting. A work session around our artificial intelligence today. I know that this is been rescheduled three times maybe four but we're here and thank you so much. Can we have roll call please. Director Abelades. Here. President Vane. Here. And we are still waiting for Director Tidholm to join virtually. We will wait or we'll we'll just have her join us when she can and get started. Director Abelades. Yeah as you said Director Vane we have many opportunities to have this work session we're finally having it. We were checking the weather report yesterday just to be sure that AI was going to be cooperative today but really excited to be able to share with you some of the great work that the team's doing with this tool
002and actually do some hands on application to using artificial intelligence. But as at a conference last week and one of the speakers who was talking around different ways that he's helping districts utilize artificial intelligence but he started off with a really great quote that I think was very appropriate and like there now is that technology is neither good nor bad nor is it neutral. So I think artificial intelligence fits in that same way that we definitely have to be paying attention because it has to have a huge impact on our students and the work that we're doing with our educators to support our students. So look forward to sharing the great information handed over to Tisha. And Director Tidholm has now joined us at 401. Thank you very much. Director Tidholm can we just do
003a quick sound test to make sure that you can hear us? Hear me? Yes we can thank you very much. Thank you. Thank you. All right so good afternoon. Grateful to finally be able to sit with you and talk about generative AI and some of the uses in Highline and joining me are my director of digital learning and library, Megan Terwilliger and Holly Fergie, Chief Policy Instructor. So allow me to take you back three years. Three years ago no one in education was talking about AI and if it was, it was in science fiction. And then in November of 2022, OpenAI released their first version, Catch Me to the Public and that was the first time where generative AI was publicly available and able to use in a lot of spaces. Since then you can't
004discuss any industry without mentioning generative AI everywhere. Which is a little bit of a difference between if we would have had this back in September, it was just starting to really get integrated but now it's fully implemented in a lot of places and a lot of other places are at the basic steps depending on the tools that we're using. So our intent in this work session is to communicate how we see generative AI use living in our system and how we intend to apply our promise and our goals towards implementation and how we. Excellent, is that better? So when we, let me go back. So generative AI fits in all four of these. You'll see in a slide a little bit later on where we actually detail how exactly it fits. I think where we're
005focusing on mostly in this space is future ready as well as we're touching on innovative learning and some of the things we're gonna talk about in the future. So we're gonna weave in our goals all throughout the presentation. And when we think about our promise, again we'll kind of land in the space of the future they choose. What we're talking about really is even changing potential futures right now. There's so much power in generative AI as well as the practices and some of the tools that are being used to where it's changing trajectories. And so we're doing everything we can to be as intentional as possible with the implementation. And then again, we're at the early stages. So our purpose in this time is really to give you an overview again of generative AI. I'm
006gonna walk through a lot of the high level basics as well as get into some of the tool use. We're gonna provide some examples of some possible uses here with the tool. We're gonna talk about how guidance and policy impact implementation. And then we're gonna give you some information that hopefully allows you to be able to engage in some policy work going forward. And we're gonna talk a little bit about this here in some slides coming up. So the main two policies that we're referring to and there will be other policies that we might touch on in the future. But policy 2022 and 2026. 2022 is commonly known as our acceptable use policy. We're working on changing some of the language to responsible use. And I think the biggest difference is acceptable use is a
007list of things that you should not do. And the responsible use is a list of things that you can do and to make possible. So we really want that type of language to be able to support our students. And policy 2026 mostly focuses on devices and electronic use, but we feel like there's a significant overlap in that policy when we refer to generative AI. So specific to the guidance and policy, one of the driving questions that we've asked ourselves, what guidance and policies are needed now to address the immediate needs? When we think about guidance and policy, really think about this idea of governance that's based on guidance. And specifically, the analogy that I've used with the team when we speak to governance is this idea of, governance will act as brakes for anything we
008try to do. And sometimes you think of brakes and just in terms of stopping, but what brakes allow us to do is go as fast as we possibly can, knowing that we have the safety of being able to stop when we need to, right? Or being able to guide ourselves into proper places. So we're gonna ensure that the AI use complies with, again, our existing digital policies and do as much work as we can to make sure that we integrate in those other spaces. So I'm gonna allow Holly to speak to some of the policy work. All right, thank you for that. Yeah, 2022 and 2026 are three, they're old, and they don't really, they don't contemplate this world at all. There's obviously some work to do in there. We did take an opportunity last
009spring to update a number of our discipline definitions to include the use of AI, like in the creation of materials that could be used to bully students. So that's one place where we came a little bit ahead of the game, but I just wanna name, we don't know what we don't know. We don't wanna write policies that restrict us to what's maybe true today, knowing that in this world, that may not be true in two days, and also recognizing that when you're at the nascent stage of using a new type of tool, it's hard to write a policy that's really gonna be reflective of what you want to be doing with that tool. So we don't exactly know what those new policies are gonna be. We know we need to update 2022 and 2026, that's
010for sure. But I took the opportunity today to utilize a tool that Dr. Brand created of a, shoot, I just forgot the word, co-pilot agent, thank you, for school board policy. And I thought, well, let's use AI to decide what our policies might look like. So I asked it a couple of prompts. I'm not gonna give you all the details, but one prompt was, is there anything that an AI policy targeted at students should include? And it gave me eight things that we need to consider, data privacy and security, bias and fairness, compliance and monitoring. So those are things that we will be considering as we think about what the policy suite is to have that supports this work. Another one was, is there anything that an AI policy targeted at students should avoid? And
011that gave me six elements, including don't be overly restrictive, don't ignore ethical considerations. Again, things that we'll be able to take this into account as we develop the policy. And then the last one that I did was, and this is gonna reflect back exactly on what Sean was just saying, are there situations where you should not have a policy around AI? And that gave me five elements. But the one that I think is really important is they said, if you're in the early stages of AI adoption, it might be more practical to focus on understanding and experimenting with AI technology before formalizing policy. This allows for a better understanding of the potential impacts and needs. So that's kind of where we are today, is in, again, that nascent, that beginning stage of adoption. Some of
012us are much further along than others. But I think this was just a really good example of how AI tools can help us create better work moving forward. And frankly, would I have necessarily thought to think about, are we a small scale operation or a highly regulated industry? I don't know if I would have. So it's just the tools themselves are giving me ideas of how to write a policy that's gonna be appropriate for students and staff, that's not gonna box us in, that will allow us to be flexible because we know we don't know what is coming. It was kind of fun too. I do have a question, and I like being able to play in the sandbox prior to creating that policy, but looking a little bit further in regards to curriculum, how
013do we go down that path and make sure that curriculum and a week later, I'll see, superintendent creates the procedures making sure that all flows in well together. I'll see, updated since 2020. So I like being able to play the sandbox. I think there's enough that needs to be different about 2022 and 2026, that it will be worthwhile to have some conversations rather than just coming forward and saying, here's the new policy, especially since I think that one of the things that I'm certainly supporting this stuff more is there really is a need to be flexible and policies, both all know, you can write a policy that ends up sitting there and really starting to box quickly. We really wanna make sure that we don't have that. This is not the enroll environment. Yes, that
014is a commitment. Thank you, Holly. So one of the things that we did was we worked as our AI group, which I'll talk a little bit about later, but we also, and again, apologies for this text being so small, we had to fit it all kind of on one page and hopefully the printout helps a little bit, or maybe you'll be able to scan it and run it through AI so we can read it to you. But there's some opportunities here just to see how the four strategic plan goals align to how AI can be implemented in our system. And I'll give you a little bit more time to read those through a little bit later. I like the, again, the future ready definition, which really explains how we can give our students opportunities to
015be able to succeed in this space effectively and ethically. And there's a little bit more discussion about ethics, which we'll get into a little bit later. So why is this important now? I mentioned it a little bit before, but AI is this transformative technology. And if we do not have this conversation now, then we will be led rather than being able to guide our own direction, because as more tools come out, they force you in a very specific direction and it's good for us to know and understand so that we can manage all the tools that we need to use as a system to your earlier point, so that we can support the direction rather than just having to follow as things move forward. So a real quick overview on generative AI. And this is,
016again, gonna be some grounding what generative AI is in helping you to understand how you can apply that and some of the work that you need to be able to do moving forward. So some common perspectives on AI. In the right side there, you have the vision of AI that is gonna take over the world and do everything for you and kind of push you in a lot of different directions, which is not what we want. The bottom left is R2D2, which is the helpful assistant, which it can be at times, but you always need to check its work. And then the one at the top is, we're just gonna hide from it, it'll pass us by. I remember in January of 2023, there was a lot of perspectives on, well, this is a fad
017and a trend, it's going to go away. But I remember the same thing being said about social media back in the day. Social media is gonna go away, people are gonna get tired of talking online and having these conversations. I think this is heading in the same direction and that's why we're trying to get ahead of it. So a few things that AI is, I think the ones that I'll point out most is warehouse fulfillment. I think previously, there were some machines that would work in warehouses, but now with AI, it's a little bit smarter to where it can do collision detection and a lot of things which allows things like overnight shipping without people involved, as well as text-to-speech, which I believe a lot of us have used and also foreign language translation. That
018has become a lot better since the early days of Google Translate. And it's in all of our lives and industries. What is not is not smarter than us. And also not only for geniuses and it's not going away. Anyone can create a prompt, as long as you can ask a question, you can interact with generative AI. The quality of the question definitely depends, helps with the quality of the response. And that's something that we're gonna talk a little bit about once we get to the prompting section. So the definition of generative AI, the things I wanna point out here is natural human language and the new content that is created. So the definition of new is what's under a question right now. Since how generative AI is built, is built off of the internet. So
019scraping the internet, getting data sets, pulling those things together. So the question is, is anything that's ever created new? And I would argue that yes, but it's just like if I was to write a poem right now, I'm using words that have been used before, but I'm putting them together in a different way. So if we take that definition, a lot of the things that we do create are new. The difference is that right now, not just a text, but also images, media, music, all types of art are created based on brand new information that's given in real time. So I mentioned earlier about prompting. Prompting as a skill is the equivalent of coding, but using natural language. So what used to happen was that in order for the computers to understand this, we needed
020to understand its language. The difference is now we can speak in any language, any language, and communicate our intent, and it will respond to us based on how it interprets our intent, which is really a seismic shift in the world. But most of the older models were literal when it came to that. So we give it a very specific information, and it respond as literally as possible. But now what's happening with the models and the model of growth is the fact that we can give it very vague terms, and it can respond understanding our intent. But the thing that we need to remember is context is king, and I'm gonna continue to come back to that context is king process. So on the left, I asked the prompt, or asked ChatGPT in this case to
021please draw a bird. And so what ChatGPT did was it drew just that bird on the left based on what it felt my intent was. On the right asked a very specific question. I wanted to draw a blue jay park bench in a city, and the blue jay would normally have a tat, and that the setting again is a park. The background is blurred using bokeh for the bird. Without parameters, ChatGPT will try to create whatever it thinks that you want, which is what happens on the left. But I was very specific on the right. So again, context is king. You'll often have instances where ChatGPT will just, it's called hallucinating. It'll give you an answer based on what it thinks that you want, but the way that the early models especially were programmed was
022to keep you happy. So sometimes it would create or make up an answer. And so when it made up an answer, that was a hallucination. And the newer models are really guarded against that. And I'm gonna pass it to Megan. So we wanna just take a moment to really think about everything that Tashawn just shared and kind of think about it in the context of education and kind of connecting it back to our promise and really thinking about how it can be utilized to develop and further each student's educational goals. Thank you. And so this slide represents sort of an early iteration from Teach AI in terms of how we were thinking about the use of AI in the context of education. Again, kind of putting in that perspective of what it is for education,
023what it isn't for education. But really for me as a former special educator, some key things that I think that have a lot of potential in some of the ways we're thinking about it and we've seen applied across in various spaces in Highline is really thinking about how do we personalize and adapt content to really make it relevant for students? So for example, a teacher can very easily and very quickly generate a text, providing context that is relevant to the student, you know, focused on interests that the student is really important to them, perhaps in the language that most connects with them and just allows for a lot of opportunity for them to really connect with the text in a different way than a different piece. So that increases student engagement and just kind of
024creates this different opportunity to engage with students. And that can be done really quickly instead of having somebody sit there and be writing that text, it can be done in seconds. Another use case that I have seen and people have really connected with is being able to create rubrics really quickly, right? Standards-based instruction is extremely important. It takes a lot of time to really think through what is that progression of learning that's needed for this particular task. And rubrics are so important to help kids really reflect on their learning and kind of know what the stage of their learning is and helps teachers really articulate what it is that we're learning, but it takes time to develop. And so that's an opportunity that I've seen really be connected to how we think about the use
025of this tool or these tools. Translation is another opportunity. You might be familiar with Google Translate, right? It's a machine translator, but there's limitations there, right? And so we have seen a lot more accuracy and contextually appropriate translation. Still a machine, right? It's still not a human translation. We still have to be diligent and review that output that is generated, but it's just another opportunity to kind of increase the efficiency of our work. In terms of what AI is not, I completely agree with what's on the slide. It is not, we are not intending to replace teachers or any educator or anyone on our staff. We are really just looking to increase efficiency so that we can really focus and hone in on our most precious resource, which is time with our students. And so
026when I think about how we integrate these tools in our system, I really think about how I think about any sort of educational technology implementation and really identifying what is the purpose, what is the need, what problem are we trying to solve, and then identify the tool and the use case that will really help us solve those problems. So those are just some ways that I think about how we can connect this to our instruction. So here's a quote that early on when we were starting this work that we really kind of centered some of the work on is AI is good at describing the world as it is today with all of its biases, but that it does not know how the world should be. Part of our implementation work is to safeguard against
027some of the shortcomings of an intelligence that was built off the internet, built off of information on the internet. And when you really think about what the internet is and can be, it's across the spectrum. There are a lot of things to rejoice in, but there's also a lot of things to definitely have some concerns about. And so it's very important for us to be able to guide, and that's one of the things that we discussed a little bit earlier. So some of the considerations that we have as we've been trying to create our guidance are these three basic considerations. And when I say basic, I don't mean basic in terms of their weight, I just mean they're just straightforward considerations. First being equity, because we want to ensure that all of our students have
028equal access. What ends up happening with new technology, especially, if students do not have access, then a lot of gaps widen, right? Our students don't have the opportunity to learn how to use the tool, don't have access to the tool, can't use the tool to supplement in certain ways in terms of having an at-home tutor and some of the other use cases that I've seen students use this with. So without that consideration, we don't go as deep as we can in some of those spaces. In terms of ethics, really the ethical portion is how do we know what decisions are made, how decisions are made, and then what are the moral implications of using AI in education as well as in our operations. And we've been focusing on operational and instructional use. And with bias
029is the use of AI to perpetuate and amplify biases that are found in data. At times, the AI might make a judgment based on what it knows and understands, back to our last quote, but there's times where we need to guide it out of that space by being very specific in the questions that we ask. So, and this just goes into kind of the same things. What we're trying to do, again, with our consideration of equity is to, in the use of AI is to preventing a digital divide in the work that we're trying to accomplish. And a couple of guiding questions specific to this, and when we think about equity and AI is access and understanding. With access, we wanna ensure, or how do we ensure all students have the access that they need
030to safe and secure AI tools? That's something that we're considering in our guidance to make sure that when we communicate this to schools, they know exactly how to support students who don't have the same capability or ability to be able to either purchase or access a free tool. And then understanding, how do we know that our students have the grounding in digital citizenship, digital literacy, information AI literacy? And the reason that those things are important is that sometimes the thought is that if you use the same exact skills you use in general, in society, you can use those same skills in the digital world. And yes, there is some parallel, but you also have to understand how to navigate some of the spaces. When I used to have this conversation with my boys, I would
031explain to them that the world that I live in is a lot different than the world that you live in. I remember my dad telling me that you can walk down certain streets, but there's always alleys that you should avoid, right? And just knowing that helped me to kinda keep safe, especially when I was out doing the things with the friends and such, and kept me to understand exactly where space and place was. It's the same thing in the digital space. There's certain sites that you should know and avoid and know that from a young age. We wanted to have our students know and understand what it means to be able to have, again, digital literacy, information literacy, and AI literacy, so they can know and see the truth versus what might not be, yeah.
032I'm just wondering what kind of training do our staff get or will be getting, but also continue? Because as you say, oh, this is evolving, right? This is just the, we're just scratching the surface and with the ethics of all of this, I know some staff are much more willing to go down this path and so much more resistance because they're just not, not that they're not willing, but maybe this is not their forte, so I'm just wondering, with any PD, we put out, we launch it, and then I know with our staff turnover or even sometimes just changing even schools or location, because we usually roll in a way, and then we start this way, and I'm just wondering, with any PD, and especially with this, with the ethics piece of it, and I
033feel like this has an opportunity to go wrong really fast, and like you were saying, going down those roads that we should not or those alleys, so the safety of this all, it was like with any internet, right, websites or whatnot, but I feel like AI can really kind of lure you because it just, I deal with it quite a bit even in my work as I'm asking it to supplement some of the things that I'm putting out, how do you even, how do you say this even better, or how do you access, and so I'm just wondering, it's a great tool, but how do we keep ethics and teaching our staff to engage our kids, but also, it's a great tool, but it just worries me. Yeah, for sure, and a very good question,
034I appreciate it. We, as a larger group, and I'm gonna mention it a little bit more, we have an AI ambassadors group who is a combination of central office staff as well as school staff teachers, we have some principals on that group, we're gonna expand to students in the near future, and those are the kind of things that we talk about. We really go through exercises on ethics so we can understand the ethics of AI more deeply and broadly and be able to communicate that back to the people that we're invested in and support, right? That's the reason we call them ambassadors. They represent wherever they're from and we teach them all the ways that they can incorporate that, and this is a group that's not going away, so we're gonna continue to engage in
035that. We also have some smaller groups within that to develop some of the PD resources in order to ensure that we have the pulls to pull from, whether or not you're a veteran teacher or a new teacher, you can still engage in those ways, and there's a couple of other ways that we'll kind of get into in the future, and did you wanna? And also just thank you, I think involving our students because they're the ones who are in it and they're gonna be probably finding ways or accessing things that we wouldn't have thought of, so just thank you. Yeah, I think I also wanna go back to what Holly said earlier because I think it's so important around, you know, as we're in the early stages of exploring AI, we really wanna just think
036about how we're experimenting, and I think that's what part of the ambassador group is helping us do, and even within cabinet, you know, Tashan's done some training for us, and we've been trying it out, you know, it's co-pilots embedded within our Microsoft tools, which allows us to be able to do, you know, coaching responses and email, be able to do different searches, and now with these agents, which we literally are just experimenting with right now, it's really opening up the possibilities for us and thinking about how it could be, kind of describe us like that. So I think it's really important that's adapting and supporting us and hopefully giving us more time to be able to focus more on some of the greater work we're trying to do. Yeah, and I would just add on
037and say, this is such a different world that we live in. It's not the same as like, you know, I supported the standards-based grading synergy gradebook launch, right? Like that's a whole like really specific bucket of learning with really specific categories of things that you can roll out and PD, and there's a whole implementation plan. This feels different to me, right? It is so much already become a part of our lives in a lot of different ways, not just professionally. And so I think that it's gonna be an ongoing conversation. And that's, I think why we're being so intentional about really developing clear guidance before sort of rolling it out, right? Because that creates that sort of space for us to say, okay, this is sort of the box without putting us completely in a
038box. And now let's have some conversations about it after we really understand the tool. And so that is why it's been so important for us to really engage in utilizing the tools so that we understand them better because we can't really make decisions without understanding it. So I think that really leads us into the next piece nicely, really thinking about the ethics and the moral implications, and anything with ethics requires a lot of just grappling and conversation and discussion and perspectives, and really thinking about how these different pieces roll out. And so really where this plays out, oftentimes in the classroom, is it tends to come, the very first question we often get is what about plagiarism? What are we doing about plagiarism? How do we handle this, right? Because that's the first kind of
039ethical thing we have to think about when we think about connecting with students here. And so this is an example from OSPI guidance that came out a while ago now that we're leaning into. And it's just really a nice example of a way to bring transparency to students around the extent to which they might or should use it across different contexts. So it just creates a good discussion piece around, okay, for this specific task, for this specific purpose in your learning, we're gonna stay in this AI-supported drafting space for right now, or maybe we don't want any sort of AI assistance. And so bringing transparency and clarity around how you might engage with this tool or how might you not engage with a tool in specific contexts just helps for that transparency. The conference that
040I attended a few weeks back had some students actually talking about their experience across their districts with us. And some of them said, nobody's talking about it. So we're just figuring it out, right? And that's what happens with kids. If we don't talk about it, if we don't bring transparency to it, they're gonna figure it out and they are smart. And so I think that our focus and our guidance is really thinking about how can we create that space in any context to have those conversations with kids and lay out clear expectations on the front end so it's not a gotcha afterwards like, oh, you used it and I didn't tell you not to, so now what, right? And that creates opportunity to really think about you know, when it isn't quite so clear, how
041do we handle that and let's talk about it. And that's definitely a piece that we're gonna continue to pack in the AIA group and we'll continue to be providing some guidance for teachers. So the third of those basis that we talked about is bias in AI. The idea that there's a need for this critical evaluation of AI outputs, meaning that there's a human portion that never gets left out of the equation. We always have to evaluate what it gives us so we can ensure that it matches the expectation, I mean, does not perpetuate or amplify any of the biases. So in bias, as you can see, two different images here. The one on the left was the original one that said, please create an image of a nurse treating a patient. And they gave us
042its representation of what it felt a nurse should look like and what it felt a patient should look like. And if you kind of look closely, you can see several different things. But on the right side, one of the things that we use as a team and we teach as a team whenever we have these AI sessions is we talk about the process of refinement. We feel like refinement is really the opportunity to be able to engage with the AI components against anything that has to do with bias. So the refinement that I used on that original prompt was recreate with an alternate representation of the nurse, but it also gave an alternate representation of the patient as well. So again, its bias is built into the general AI datasets because it was trained on
043the internet and refinement is necessary. And again, context is king. So we ask ourselves this question when we kind of think through those three things, how do we support equitable, ethical and non-biased AI use in Highline? So Washington State put out some excellent guidance, I believe in the beginning of 2023, which is a human-centered AI guidance for K-12 public schools. A lot of other states have adopted variations of this in order to move forward because it's so solid. But what is built on is this idea of human input, which is the prompt, the AI output, which is the response that it gives, and then the human interaction to determine how well that response was. And so the point that I'll make here is that subject matter expertise will never go out of style, meaning that
044we will have to have people who know things in order to use generative AI. Because I've heard a lot of pretty wild statements about, oh, well, reading is gonna go away. And I always kind of go back to, I believe it was 1999, where the IBM supercomputer beat Kasparov, the world chess champion, and his statement on the day after was chess is dead, right? That was 1999. I believe people are still playing chess as long as I'm aware. So some of those statements are really overblown, but there's an expectation that subject matter expertise will never go away because people have to know things in order to use this properly. So generative AI use in Highline, these are the principles and values that we've used in order to create the documents that you have in front
045of you, the vision, the instructional, and the operational guidance documents. Human-centered, safe and secure, equitable and inclusive, purposeful and beneficial. And to walk through those quickly, human-centered just means that humans are not left out of the process and should never be left out of the process. Safe and secure, we practice digital citizenship, wow, hardest word ever to say, digital citizenship standards and protect information and self. We help people to understand how to protect themselves while using the tool. And we make sure that we also protect information and data from getting out into the hands of third parties. Our AI practice considers and accounts for all. There's no aspect of our system that is left out of our AI implementation. And we're making sure that we engage with all the people who exist in all of
046the spaces in our system. And we're not just doing AI as a trend. We're really trying to make sure that we stay ahead and really focus on how do we move and give our students the opportunity to participate in the future. And so the ambassador group, which I've mentioned a few times, again, their purpose is to implement AI within the district aligning to a lot of their roles and responsibilities and being able to give input, not just on their roles, but all of their work. And again, as principals, teachers, administrators, specialists, and some students, and I believe we have some of the people from that group here. And so just wave or something if you're part of the AI. Cool. So yeah, here to support. And one thing I'll point out is that any image
047that, and I use this for all of my presentations, any image that I use generative AI to create, I always put the prompt underneath the image. So any of the images that you'll see, you'll see the exact prompt that I used. I encourage you to try to use that same prompt and see that you'll probably get something different, which kind of tells you the fact that you have to have a little bit of discernment whenever you do these prompts. So again, the three documents, or yes, the three documents that you have are the vision document, which is the high level, what we see as the purpose and potential, the instructional guidance, which is to help classroom staff, as well as school staff to be able to navigate a generative AI use in schools. And then
048our operational guidance, which really focuses on our best practices for how we use this operationally and all the other aspects of the work that we do as a system. And again, each one of those is in draft form. They're actively being updated and they will be living documents as we move forward. And so I wanted to make sure that I included some of the principal specifics to the operational side as well, because we did mention the instructional. And so data security is really important. A lot of the third-party tools actually take the data and train off the data. One of the reasons that we primarily focus in the Microsoft Copilot space is because the data is contained and is protected and we have controls over any place that the data goes. So it's very structured
049and protected. We're focusing on training and development to your question earlier, and this will be consistent and ongoing. And as we do continuous employee training across our spaces, we're trying to do at least one training every month or two months to make sure that people are aware, but that's across instructional and operational. So we might do an instructional one at the same time as we're doing an operational. And then one of the things that we've talked about is we need a method of transparency to ensure that we are doing, whenever we use AI, we're being transparent with the fact that it's being used or it helped to co-create information and we did review. But there are some times where we've had conversations on, is this any different than Googling something? Because people don't always put
050citations when they Google. So we're trying to figure out where the balance is. And once we see some good guidance or create our own good guidance on this, we'll be transparent about what that is. Well, actually, no. This is where we get to our hands-on portion. Sorry, I went through the last part pretty fast because I wanted you to be able to get hands-on on some spaces. So in your browser, if you have your laptop, yeah, play too. So yeah, you can go to copilot.microsoft.com. And if you have another AI assistant that you use, you can use that as well, since we're not using any protected information that will work just as well. So just let us know when it helps you get there and then we can help with the rest. As long as
051you're logged in, it shouldn't make you. All right, so once we're logged in, I think what we're gonna do is AI for document review. There's a couple of documents that you have access to that believe that they are state bills for special ed and I think MSOC funding, or the two. Yeah. And director to home, are you able to access the access copilot? Yeah. Well, I'm using chat GBT because I have an account, but yes. Okay, perfect. It'll work the exact same way as long as you have access to the document. So a couple of quick steps, if you're already logged in, I can kind of move on from here. All right, so a basic prompt that you can use is summarize the attached document. I want you to give that a try and tell
052me what it tells you. So the SB bills, I believe we're attached with the PowerPoint. We can help get those to you if they are. Yeah, or we can email them to you. Yep. Excuse me, Friday comms. Yeah, 52, 63 and 53.07. And I believe 52, 63 is the one that's still capable. Director to home, we're gonna mail you a file. Okay, thank you. Yeah, we sent them the first time we scheduled this and I believe the second time we scheduled this and didn't send them the third. I just found the files. I found them in the Friday comms. Okay. And 53,07, but we can just do 52, 63. Yeah, once you have the file, you don't wanna put on your desktop and documents folders and upload it. Yep. So put the steps back up
053so that you can see, you can save it either on your desktop or your documents folder. Yep. And then once you get there, then this is a part where we can help. So once he has the files now, you already put it in. Summarize the attached document. So one thing that you'll notice is that when you summarize, when you do just summarize the attached document, it gives you some basic things. It sends you in the direction that it feels that you should go. And sometimes that can be problematic because there might be some things in there that you really need to know and understand. And so when we talked about refinement before, that second prompt, the please provide, gives you a little bit more detailed information. So give that one a try. And yeah, we're
054bringing it up on the presentation screen so that everyone can see it either online or locally. So Director Tidholm, were you able to run each one of the prompts? I think it's finally working. It wasn't liking the format initially when I uploaded it. So I did just get a summary, yes. Okay, perfect. So live demos are always fraught with peril, as most people know. But we did get some responses for an initial summary as well as a little bit of a deeper prompt. And once the prompt refinement happened, question, were you able, Director Vann, to get the second prompt? And what were the results that you got? Could you describe them, please? Yep, the second one. Was it more closer to what you would expect in terms of the things that you're looking for? Yeah,
055yeah. So one of the dangers that you have is even when you get the initial response, sometimes it can lead you down a rabbit hole. It's kind of like watching your first YouTube video and then 30 videos later and you're in the space. Same thing happens with generative AI. The analogy that I'll use is that oftentimes it's kind of like when you're whittling a piece of wood, you can't put the wood back on. So sometimes you get too deep. There comes this part where you probably need to stop or start over again if you're not getting the results that you need or expect. Bring it back up. All right, so thank you for engaging in that. Hopefully you were able to see kind of. Yeah. Hopefully you were able to see, one, how easy prompting
056is, but two, what the engagement looks like. Got 307 from the filters. Yes, the filters. Yep, so again, the basic difference is when you ask it, give me something, then it's gonna give you whatever it thinks is possible. When you refine your prompt and you're as specific as possible, it will give you something that's much closer than what you expected. And there's a lesson in there about specificity as well as detail. And again, subject matter expertise will never go out of style in context as King. So our next steps are, our hope is that you can continue to support us as we move forward with some of the policy work. We're gonna continue as an AI ambassador group to work on the guidance and best practices. Ongoing district PD on AI, which the question was
057asked earlier, and we're gonna do that as much as possible. And then stakeholder communication. We're gonna make sure that everyone is aware of what we're trying to do with AI, and that includes the community. We're gonna be very engaged in ways to help people, to one, understand what's happening in the system, but also to provide some opportunities for training outside of our system to be able to help people engage so that they can support their students in various distinct ways. So if you don't remember anything from this session, remember that the concept that Texas King, because that's of utmost importance, and also a refinement helps us to get past the limitations of the generative AI model. Again, this is ongoing. Everything that you have in front of you are living documents, and we will continue
058to engage with you about the policies as well as our forward movement with this. So a couple of potential discussion questions here. What questions do you have about generative AI use in Highline as a result of this conversation, and is there additional information you need on the future policies and guidance that supports this work? Well, thank you so much, Tashawn. Directors, I'm gonna start with you. First, Director Titholm, any questions? No questions, but I do appreciate this work. I have heard of ChatGBT for a long time. Like you said, kind of just thought it was gonna go away. My kids have made fun of me and told me how old I was for not understanding or using it. So when I finally started to play with it, I was baffled at what it actually is
059compared to what I thought it was. And like you said, if we don't get ahead of it, they're gonna find a way to use it anyway. So to be able to try to find a way that they can use it in a way that is responsible and appropriate, I'm really excited about that, to have this tool. I wish we had this when I was a kid, but we're here now. So thank you, I appreciate your work and your presentation. Thank you. I did actually ask it. What I was thinking more about is like, again, this is just the beginning of all of this and I know our community and our families are always wanting to know more as well. So how do we keep our community, our families, also informed about how are we, not
060just using this, but how are their students, are their use in this? But I think especially, I'm just thinking about it, especially our families who, English is not their first language and I know a couple of families who already struggle to help their students, to help their students in their education and so I'm wondering, this could be a great tool. I don't know how, if AI is really bilingual or not. I mean, we could find out, right? But how do we use this tool as well to support our families in that capacity? So I'm just wondering if maybe we could use some families in the ambassador to help. Yeah, so I've met recently with the FACES team. I'm gonna really talk about some potential strategies, shortcomings, how we can move forward to start to create
061a plan, right? So we had our initial conversation, we're gonna have several conversations to come up and that specific question about translation and language use specific to AI came up and we were able to display the fact that in real time, you can ask any of the agents to speak your language when you're engaging with them. So one of the things we talked about was a potential prompt library written in native languages that you can just copy and paste in and then at that point you're engaging with the generative AI in your native language. So there's opportunities with that as well as some other spaces. And we've also talked about how do we engage directly with schools because I know that there's some, there's some partnerships that already exist in terms of family engagement between
062schools and the system at large. So we're gonna navigate those as well in order to ensure exactly what you asked for is in place. But on Monday we had our Family Action Committee, we had some of our team is here that supported that but we broke into groups that night and helping our families understand everything from budget to our four priority areas, MTSS, literacy, secondary school redesign and inclusion. And we also had a budget and family engagement. And so I was just, we were here in the room and for me it's one of the first times I think that we're really helping our families understand the work that we are trying to do within the organization. So I can see this as a great opportunity especially in the area of literacy and future ready, the
063secondary school redesign work that we could lean upon our family actually can help us dig into this a little bit more too because they're really excited about being able to learn some of this and they already provided us with great guidance and some, the language that we use and how we can help them better understand that's gonna be very useful. I think it's an opportunity even for us to pilot that there around, you know, start building that language into AI Family Action Committee that's important in that. And I think also I'm thinking about, I know right now that we need to continue to build that trust within our families and our community and I think it's an opportunity for us to like building the co-creation of all of this because this is for our families,
064for them, by them and having our students and our families build that language because like you say, I mean, even Weiwei, when we started computers, when I first started in school, I did start with typewriters, so I am that old but thinking about, you know, when computers first started running reports, it's like, you only get out of it what you put into it, right? And AI is the same thing. You only kind of get out of it what you put into it so I think begin co-creating the languages or the information that's gonna be there, kind of same thing. Thank you very much. Thank you so much for the presentation. A couple questions here. On page 22, the scale for students in regards to what level should we be challenging our students? I am dealing
065with this at home since their phone, their Google's at their fingertips and they can come up with something like that is how do we challenge our kids to use their minds and thoughts first prior to them even hopping on the interweb and having that easy way out and how do we truly challenge that? I wanna make sure that our students are engaged in the classrooms and it's not an easy way out so that's something that I'm trying to figure out as a parent, I'm using a lot of this in the private industry AI to support what we're trying to do and it's amazing what access we have at our fingertips just because it's evolving. Every single day it's always evolving, grabbing stuff from the internet so how do we, how are we going to challenge
066our kids to use their minds first prior to jumping to their act? Yeah, I appreciate that question. For me it starts with how we think about and the types of tasks we put in front of kids, right? That design of instruction that really thinks about, okay, what do kids have access to? How are we going to get them to engage perhaps with each other first, right? It starts with designing tasks that aren't so rote, right? That aren't so direct or that really require reflection or personal experience, right? Things like that that can't come from a generative AI, right? It has to come from the person and so really rethinking and I appreciate Director Alvarez's comment about connection here to secondary redesign because that could be an opportunity to really think about what does that look
067like? Well, thank you and then the last question is how are we going to measure the impact on reduction? You were talking about creating rubrics, it's gonna help make those come a little bit faster but how can we measure that impact and say, you know what, we're gonna help, we're gonna save X amount of time or what? But also student outcome. Yeah, that's a great question. I didn't ask that question. So yeah, a good portion of that might be qualitative because what we've talked about in our group is again, to your point, how do we measure this? Part of the reason that we have AI ambassadors is because we want people who have relationships in their environments to be able to ask the question, what impact have you seen? Because if we were to have
068some way to directly measure that, we'd probably already be measuring that in other areas and it just doesn't exist in that same way because efficiency from person to person is different, right? It's just kind of challenging to measure. So we have to hear stories, we have to sit down with people engaged and say, what questions did you ask and what was the difference between the last time you did the same task and now, right? And so as part of evaluation and program evaluation, we can get deeper into what some of the successes were and then be able to tell that story. Thank you very much. All right, well, thank you very much. It makes me also think of the speakers I heard last week also talked about when we brought washing machines in as it
069was a new technology and that everyone's claiming, oh, this is gonna save us tons of time, it's gonna be great. But actually what the washing machine did is that people just spent more time doing laundry. So I think there's the same with this tool, we have to really think about how we're gonna use the most effective ways. I really wanna thank Tiffan and team, especially ambassadors work and everything you're leading there because we're really getting insight from educators and thinking about how we can use this in a much better way and I think there's a lot of possibilities and I think as we've had conversations around the power of public education and we've talked about how important it is with diversity, equity, inclusion, accessibility, I mean, this is one of those examples where we still
070know there's digital redlining, we're not all of our students have access to Wi-Fi at home and we do not want our students to miss out on this technology. So our ambassador group and Tashan and his team are gonna be really important to help us guide us through this work. We're piloting a tool right now called MagPy which is a kindergarten tool that has an artificial intelligence in the background, so we're planning a couple schools just to learn from it but it's definitely gonna be having an impact with us and we wanna make sure and be ahead of it and work with it in a way that will support our students and our educators as best as possible but thank you all for the great work you're doing. Thank you so much. With that, that concludes
071our work session. Can I please have a motion to adjourn? I make a motion we adjourn. Okay. All in favor? Aye. Aye. Any opposed? That's it. Thank you everybody. Have a great evening.