AI and the Future of Personalized Education with Edupreneur Amanda Bickerstaff

Explore the educational revolution where AI becomes the teacher's ally, not the adversary. Unlock the promise of a tailored learning journey as we unpack the integration of artificial intelligence into the classroom setting. With insights from the Houghton Mifflin Harcourt Educator Confidence Report and the expert knowledge of Amanda Bickerstaff, a former high school biology teacher turned ed-tech CEO, our conversation navigates the landscape of technology-enhanced education, from the practical reduction of teacher workloads to the pioneering role of generative AI tools like ChatGPT.

Discover how AI shatters myths of omnipotence, serving instead as a powerful computational partner that thrives on precision and rich data sets. We scrutinize the role of AI in fostering academic integrity and debate its impact on the timeless issue of dishonesty in education. Delving deeper, the episode underscores the transformative potential of AI for students with disabilities, offering personalized support that transcends traditional learning barriers. By emphasizing the significance of AI literacy training, we advocate for a balanced approach to technology, ensuring it strengthens rather than replaces student creativity and intellect.

Join us as we illuminate the ways in which AI technologies are revolutionizing support for disabled students, providing barrier-free communication and the freedom to manage their educational endeavors. We celebrate the triumphs within our educational systems by inviting listeners to share their own success stories, fostering a communal spirit of innovation and inclusivity. This isn't just a glimpse into the future of education—it's an invitation to participate in shaping it. Your stories, insights, and experiences are the keystones to this growing educational edifice, and together, we write the next chapter.

To learn more about AI for Education, go to https://www.aiforeducation.io/.

Transcript

Dr. Lisa Hassler: 0:10

Welcome to the Brighter Side of Education. I'm Dr., Lisa Hassler your host, here to enlighten and brighten the classrooms in America through focused conversation on important topics in education. In each episode, I discuss problems we as teachers and parents are facing and what people are doing in their communities to fix it. What are the variables and how can we duplicate it to maximize student outcomes? Today we venture into the dynamic realm where education meets technology. How can AI effectively and ethically be integrated into classrooms to elevate student achievement? Ai holds promise for personalized learning experiences and innovative classroom tools, offering the potential to revolutionize education.

Dr. Lisa Hassler: 0:52

Unesco acknowledges the transformative potential of artificial intelligence in addressing key challenges within education, revolutionizing teaching and learning methods and advancing the goal of achieving inclusive and quality learning for all. For a deeper look as to how educators use and feel about AI, houten Mifflin Harcourt released Part 1 of its 9th Annual Educator Conference Report Outlook on Teaching in AI. This survey, conducted between May and June of 2023, gathered insights from 1,000 K-12 classroom teachers and over 200 administrators. The findings revealed a varied perception of AI tools among educators, with 12% of them finding very or extremely helpful, 44% considering them somewhat helpful and an equal 44% deeming them not helpful or not at all helpful. Moreover, the report highlighted a prevailing lack of confidence among teachers in effectively using AI tools. While a majority of 57% recognize the importance of AI utilization and an even larger majority 58% expressed a desire for additional instruction, only 1 in 5 felt adequately prepared to integrate tools such as chat GPT into their classroom practices.

Dr. Lisa Hassler: 2:14

As we explore the role of AI in education, we must balance the potential benefits with ethical considerations. The rapid advancement of technology often outpaces regulatory frameworks and policy discussions, presenting significant risks. Join us today as we delve into insights, dispel myths and uncover practical strategies for harnessing AI's potential to empower educators and support students on their educational journey. Today, I'm excited to speak with Amanda Bickerstaff, a former educator and visionary founder of AI for Education, a company dedicated to leveraging AI to enhance education. Welcome to the show, Amanda.

Amanda Bickerstaff: 2:54

Oh, thank you. Happy Monday.

Dr. Lisa Hassler: 2:56

What is your background in education and how did you get into artificial intelligence? That's a great question.

Amanda Bickerstaff: 3:02

So I started as a teacher. So after college I was a high school biology teacher as an alternative route, which means that I got my master's and learned on the job, which is pretty intense. So I taught for a couple of years, have a health issue, so I wasn't able to do a lot of time in the classroom. If you've been a teacher or are no one, you know it's a pretty taxing life, and so instead of staying in the classroom, I started working in everything around education. So I got most of a PhD. I am a PhD dropout. For everyone out there that made it, I'm really proud of you. My grandmother is still upset about the fact that I didn't but kind of took that research background and then also started working in education technology. So I spend a lot of time working in education technology both in the US and also in Australia.

Amanda Bickerstaff: 3:46

So my last role I was the CEO of an education technology company in Melbourne, australia, and you know it was really funny. I was, you know, first time CEO in a new country that I'd never been to right before COVID, and Melbourne was the most locked down place in the world, and so it was a really like trial by fire, let's just say that. And so you know, spent three years, learned a lot, published some major research on the impact of COVID and also on student wellbeing, and then kind of took some time off. So you know, I managed to get to a pretty significant level of burnout. But actually when Chat GPT was released, I was in Japan. I was definitely thinking more about ramen than I was about for GPT and generative AI. But I came back to the US and I started to start navigating my next path. I think I had known. One of the reasons I took time off is I think I knew it was time for me to go off on my own and to start something. I think sometimes we have to be almost like the world has to help us do that, because I didn't right away start building something, but I started to working for other people and starting to work with a startup.

Amanda Bickerstaff: 4:46

Then, the first time I used Chat GPT, I asked it to build me a rubric, which is quite funny because I haven't written a rubric in about 10 years. But I know how hard rubrics are. They take a lot of time, they're super important but we often miss them because it's so tricky to get them right in the formatting. I asked Chat GPT to build a rubric and when it did in like 10 seconds and formatted it as a chart, I knew I knew that this is transformation that we've been talking about for a long time, but also that this is not an intuitive technology. This wasn't a step forward and like better Google or better lesson planning or better technology. It was something that was completely different. I started AI for Education as a website with a prompt library that you can use on any chat bot. That was back in April. We had our first school by June and now we've worked with trained about 60,000 educators from across the world since August on generative AI, and I get to do this work every day.

Dr. Lisa Hassler: 5:37

That's amazing. Could you share some practical examples of how teachers can integrate AI technologies to streamline their workloads and make their jobs more manageable?

Amanda Bickerstaff: 5:47

Right, we know the teachers spend upwards of at least five hours a week on planning and preparation outside of the classroom. We can go up to 20 hours, depending on what you're teaching. I was joking that our days as teachers don't start on Monday morning, they start on Sunday afternoon. Right, like our brains already there. We're thinking about it. The best first place is to think about your productivity and ways in which you can start to streamline things like lesson planning with generative AI.

Amanda Bickerstaff: 6:14

I talked about a rubric, like maybe you wouldn't have time to build a rubric for this new assessment, or you definitely maybe wouldn't have time to build a rubric and then a checklist and a tip sheet. What you can do now with generative AI is some good prompting and a little bit of resilience is get some really high quality, dedicated, targeted lesson planning materials for your classroom. You can also do things like helping you schedule. I mean, you can think about ways in which you can use it as a thought partner on your productivity too, like it doesn't just have to be an outcome. It could be like okay, so I tend to spend a lot of time on these five things. Let me use chat to be a tool to help me organize. Okay, so now I'm going to start a time block or I'm going to find efficiencies in the work I already do.

Amanda Bickerstaff: 6:57

And actually, something that we don't do very often as educators because we're so overwhelmed, is that there are some ticks and tricks to be able to find that productivity.

Amanda Bickerstaff: 7:05

That's what's great about AI is that it can be like your lesson planning, support your assistant, you know you can even be a mentor teacher in the sense of it can help you extend the way you're thinking about your classroom and the way that you teach it.

Amanda Bickerstaff: 7:18

But also it could just help you as a thought partner or even our organizational partner about what is the order, operations of doing this work, so that I don't feel so slammed by the end of that week. So I think that, like, honestly, what we see is that there is so much upside and potential of these tools, even though they're very new and they're the worst they'll ever be today. Right yeah, we remember the iPhone back when it was your iPod, when it first started our dial-up. These technologies take a while to catch up in terms of their reliability and their ease of use, but at the same time, chat, gbt and other tools right now can do so much for teachers to help with all those things that take up time away from what we care about the most, which is our interactions with students.

Dr. Lisa Hassler: 7:59

Absolutely.

Dr. Lisa Hassler: 7:59

I know I've touched on it a little bit here and there and I love it Just in the small ways that I've figured out how to use it.

Dr. Lisa Hassler: 8:07

It's really been impactful for me so I can see, as a teacher, how things like making a rubric I actually just asked ChatGPT to make some rubrics for me a couple of weeks ago and I was so impressed with how well they turned out. I think have to be very specific and you're prompting we kind of need help with, like how do we prompt, what are the key words to use, and then tweak it after you get that result to know. This is a really good base and now I'm going to maybe enhance it here or there or customize it, and so I just found it very invaluable for what I'm doing and I can see where teachers if you haven't been doing it so far this is. It's such a gem. It really is. It's like having a little personal assistant sometimes. I think it's a wonderful tool. What are some misconceptions that you've encountered regarding AI and education and how do you address them within your company and in your interactions with educators?

Amanda Bickerstaff: 8:58

So there are a lot I mean we have. You know it's called artificial intelligence, even though it's not intelligent. And it's really quite funny because the idea of AI, you know, even technologists up until the last couple, you know, like last year and a half, kind of didn't call it AI anymore, they called it machine learning or they use like the smaller terms, right, Because artificial intelligence, you know it, essentially is implying that these tools are intelligent instead of being. You know, these computer programs that are going to mimic human ability in certain areas, and so I was thinking about this yesterday that we really kind of we're not talking about AI, the sense of what we're doing now.

Amanda Bickerstaff: 9:39

So the first thing is that artificial intelligence is not intelligent. It is not thinking, it is computing. In the case of generative AI, it's creating. Essentially it's predicting the next word or the next pixel. If you're looking at an image generator based on all the data that it's been trained on, which, in this case, is the Internet, and then it's using your context you talked about the specificity of your prompting Like the more you ask it, the better way you ask it. It actually ends up giving you something better on the output, because it takes its training data and it's the way that it's created, relationships between words and the way that you prompt it, and then that's what you get. That's what you're essentially acting as computer scientists, using natural language or English when you're working with generative AI.

Amanda Bickerstaff: 10:16

So there's a couple of things. Number one is a again not thinking these are computer programs. They are fantabulous. I mean the sense of like they're crazy, they're interesting, it's like I'll make up a word. Someone was talking about the fact that you know, it's like having a horse and buggy to a car and I'm like it's more like a spaceship that like sometimes takes you to outer space and sometimes takes you to Jersey North, into Jersey, my mom's family is from there, but like it is like a kind of wild thing, but it is actually.

Amanda Bickerstaff: 10:43

Again, it's a computer programming that can have some. You know, it requires a whole bunch of training and a lot of money and a lot of compute power and then your ability to use it right. So I think those are two big ones. I think that also there's some misconceptions and AI is just for cheating. I mean, you know, I think that we have so collapsed this idea that now that we have, you know, a calculator for the humanities that all of a sudden all kids are going to be using these technologies only for replacing their thinking on a activity or homework and assessment. And there has been research from from Stanford that shows that you know, between 1680% of students already cheat. There's some cheating behavior that happens, and when they went back to schools this last year, they did not find that now it was 100%. What they found is that kids are cheating in a different way, and so I think that this is something that like there's there are bigger questions here, but the students that we know that are using these tools sometimes are those with special needs, like special disabilities. They have 504 plans or IPs students that are using it for help with their own studying, to get feedback directly.

Amanda Bickerstaff: 11:45

We had a student on a panel that talked about how it was really struggling with economics, and so he had it explained. Chashbhuti explained it to him like a monkey, and so it gave him an answer not in dollars, but in bananas, and it worked for him. So this is we're talking about how our kids are amazing, unique. You know people, but so I think that there's this idea, though, that we're collapsing and we're we've created so much uncertainty and so much fear about this idea of that students are going to stop learning and it's only going to be for cheating.

Amanda Bickerstaff: 12:13

But what we find is that if you give students first some literacy training especially if they like we think that's very, very important to understand what these tools are and aren't again not thinking they make mistakes, they they are going to have bias, they're going to maybe not be that safe in terms of your privacy around your student data.

Amanda Bickerstaff: 12:32

It's really important that you do that, and then people kids will start to understand that these tools need to be used as a tool and they need to be critical use or something. And when they do that, then you start to see, though that now students can you know as academic integrity is extremely important understand where the line is and be able to be transparent about their use, but then also start to think about and partner with these students around, like, how can you actually move forward together to create some augmentation of your own creativity, your thought processes, your unique powers, instead of a replacement of your thinking? And I think that it requires us to get away from this rhetoric of it's only for cheating and also to create time for teachers and leaders and students to be able to work together and learn together in this really important moment of time.

Dr. Lisa Hassler: 13:16

Yeah, thank you, tush, and a lot of very important things there. And that is there was always a certain level of cheating in classrooms. That's always been there. I remember back in elementary school we had like those pens that could like put paper in and then like twirl it, you know, like because they used to have stuff in them. And so people were like I think I'm going to put all the answers on the tiniest as possible I can probably probably good thing, I had good eyes back then, right, and they were like right on there and then you're like twisted and find it. You know, I mean, they can get super creative, let's face it. And so there's always been a certain element of those who feel like they really need it to be able to pass something important, that they're going to do that anyways, or with homework or writing assignments. So I'm glad that you brought that up and that data was out there to support that. It's not as though everybody is running to this now saying, oh my gosh, open the floodgates, we're all going to cheat. I think that there's a certain amount of integrity that students hold within themselves and so we're not going to all run to the cheating path, but it can be used as a tool, like a calculator, to say how can I use this to enhance my learning, to get me to a point, to maybe solidify a data that I need, or to make something quicker, faster, easier, so that I can move on and then add my information to it, like scaffolding. And I'm excited that you mentioned also students with disabilities. I know we're going to get into that in a moment. I find that so, so, very important.

Dr. Lisa Hassler: 14:32

I was at an educational conference over the summer for higher ed and that was the only thing that anybody could talk about was what are we going to do when our students come walking back in through these doors in a few weeks and I know they're all going to be cheating Like what? What policies are we going to put in there? How are we going to write it in our handbooks? How are we going to prevent this from happening? Well, now, here we are, almost a year later. We're getting through the school year, a semester is behind us, and I think that a lot of teachers may be kind of like oh, like a relief to say, okay, it wasn't as terrible as we thought that it was going to be. They didn't take this and replace all of their work. So, from your experience, what do you see as the biggest challenges in implementing AI and educational systems and how do you suggest educators and administrators navigate these obstacles effectively?

Amanda Bickerstaff: 15:21

So I think the first thing is this we talked about that spaceship that suddenly takes you to a place you definitely weren't explaining on going. These tools are really new and one thing about this is these technologies were not developed to be used in classrooms and with teachers and with students. The Chat GPT 3.5 was an experiment. It was released into the world as a conversational AI, a conversational framework, so it's a chat bot because it was the easiest framework to put it out into the world. So there's no reason it has to be. The model has to be conversational, but it was a choice that was made. That has like long term effects. I love how New York is loud. I don't know if you could hear that in the background. It gives you the flavor of our life on the upper east side of New York. But yeah, so we have these technologies that are not really responsibly made, so they're not particularly transparent. They have lots of problems around hallucinations we talked about, which is what this idea of inaccuracies, bias. You might have seen in the news, gemini yesterday, which is the Google, gemini replaced Bard as a chat bot and these text image generators are super, super biased and the way you get around it is you essentially pre prompt the thing to be diverse and its output. And what happened is, you know, it gave women in color and black Nazis when it talked about historical figures from the 40s. And so you know, it's so trying so hard to not to not be biased that it became significantly more biased. And so that's all to be said is that these tools are very, very early stage. They're black box models. We don't really know fully how they work, including the people that are, you know, paid to develop them and to run those companies. So that's one component of it. The second component of it is that these technologies are quite expensive, and so if you're going to be thinking about integrating some of these tools into your schools, it comes with a significant cost, especially whether you're doing it through GPT for licenses, which is going to be the frontier model, the best model on the market, or you're building your own system, or you're working with an application that is using FNI or or Google's models underneath it.

Amanda Bickerstaff: 17:21

So we have some like interesting things from the technology side, right, New technology, unreliable, biased, very expensive, and then what you have, as well as like there's a real lack of literacy about what these technologies are. It's not like this is better. Google is actually worse Google in some ways, because the free version isn't connected to the internet. So there you go. Definitely not better Google. What we see is that, like, because it's such a big step change, it's not like we could take what we've already done in technology and suddenly we have, like, the next frame where we understand what to do next, and so what we have to do is we have to take a step back and build literacy around these technologies and so that we understand what they are, what they are, not, how to use them, capabilities, limitations Because when you talked about all these schools that were saying, okay, when you went to that conference AI, guidance, you know guidelines, etc.

Amanda Bickerstaff: 18:08

What probably happened is one or two things what is a never got to it because it was moving too quickly or they did potentially wrote something that is not possible to enforce. Because instead of saying generative AI, I'm like like you're saying AI and 86% of us use some sort of AI every like week, whether it's in our phone or our laptops, and that's across the world, and so like we can't stop kids from using AI because they wouldn't be able to use Google search, for example. Oh yeah, but we can say it's like that. You can see that the literacy component is so important that even those schools and systems that are trying to create policy may not have enough knowledge about what these technologies are to be able to even write a policy that is effective. So you can see that there's a big, big change that's happening here, and so what you know this is a huge change management process that requires quite a bit. With the knowledge I mean, my life would be so much easier if I can go into a system and say, hey, here are the five tools you can use, here, the five guidelines.

Amanda Bickerstaff: 19:03

You can go, instead of being like, hey, we can help you think about ways in which to use these tools ethically.

Amanda Bickerstaff: 19:09

We can train your staff and your leadership and your students on AI, literacy and understanding what these tools can do and cannot. And then it has to be something that we move forward together, that it needs to be this deep learning process that allows for us to continue to develop these you know ideas, these mindsets, these policies, and with the hope that then we're in a position to actually change school for the better, because our existing model of education isn't really fit for purpose when you have a new thing that can do like 90% of all assessment better than the people writing the assessments. So all that to be said is that it is a pretty tricky process right now. But there are, like, if we think that the two things that are foundational are building AI literacy and then creating some guidance, that's more about mindsets and appropriate use and safe use, and then those two things happening concert. If you can do that, we see a lot of really positive moves forward and try and starting to, you know, kind of just transform what school becomes.

Dr. Lisa Hassler: 20:06

So, with the rapid advancement of AI, how do you ensure that technologies your company develops are ethically and responsibly deployed within educational settings?

Amanda Bickerstaff: 20:15

Yeah, I mean it's it's the most important question we can ask right now. It really is. I mean, there there are lots of, you know, possibility for this stuff to do great good, but also it could be do great harm, and we're going to see that and the election this year. So we're going to see misinformation and disinformation. That's going to be easier to do than ever before. We're going to see so many impacts to our lives. I think we we kind of assume we call Jen Alpha and Jen C these digital natives, but being able to use phone and social media is not digital literacy, and students fall for scams. Young people fall for internet scams more than those in their 60s and so you know, because there's our footprint online, so much greater right, and so the idea that this is an inflection point that we are living through that requires us to take a moment, take a step back and invest in building this literacy, I think is by far the most important thing we could do right now.

Dr. Lisa Hassler: 21:15

So I want to take us back to how AI can support students with disabilities. Can you elaborate on some of the ways that AI technologies can be tailored to provide personalized learning experiences for students with diverse learning needs?

Amanda Bickerstaff: 21:29

So the first thing is that something like a Chat GPT. If you use it on your phone, you can talk to it and they can talk back to you so you never have to write. And so students with dyslexia, low to no vision, students with aphasia and other things that make it hard to write this is a great opportunity to absolutely lower that barrier to getting your thoughts out. We can see it through technology, assistive technology that's being used. That use generative AI and also just deterministic or classical AI is that you know to not have someone listening and saying, oh, you spelled that wrong or like, what are you doing or you don't quite know what you're doing. To do that without judgment is an enormous way to unblock those students. They're things for students again, with low no vision. You can upload a GPT-4 and other technologies and image and have it actually explain it to you. So instead of having to have an aid kind of talk you through an image, it can speak to you. And actually GPT-4 vision, which is the new model it was first deployed with Be my Eyes, which is a low to no vision organization supporting people around the ability to have now this great on demand way of explaining images in natural language. So I think there are a lot of pieces like that. We're going to see more and more tools that are going to be developed and are being developed around personalized learning and accommodations, but just those, even these free tools.

Amanda Bickerstaff: 22:47

Now, just to be able to have these unblockers and I'll call them an unblocker. We've had assistive technology. It's all about assisting, but this idea of kind of removing these significant barriers that we try to but we tend to not be able to because we can't even one to one, it's very difficult to do that but another technology can act as that. One to one and on demand. And that's a big thing. Like on demand support, like I don't have to wait, like if it's 12 o'clock at night and I suddenly have a you know, I'm a student with dyslexia and I suddenly have a flash of brilliance, but I don't have my aid there. I don't want to know what to do, but I can actually kind of dictate my thoughts and have it make sense of it for me and in a way that I don't have to spend so much of my cognitive load just to figure out how to spell something is an example of that on demand and block it, but can happen with these tools.

Dr. Lisa Hassler: 23:35

I'm so excited about that portion. I have a son that is on the spectrum and language is just very difficult. So when he went through school, it was always the language component that really held him back, and so even just the thought of you know higher ed, wanting to get a college degree and that one thing really blocking him from being able to move forward and I think about other students as well, where they have that that one area that if you know something like Chat GPT or an AI assistant, that would open up a future that has been withheld for them. I just think it's very, very exciting and I think it offers a lot of hope and society can really benefit from all of this, and so I think it's going to really lead to wonderful things for them in particular. Thank you for joining me today to discuss AI and education and how it can be used effectively.

Amanda Bickerstaff: 24:27

Of course, yes, and thank you so much, Lisa. I appreciate being here and we have lots of great stuff on AI for Education. So you know we have a webinar coming up for students with disabilities actually using technology that I think you might find useful, but just really appreciate your time today and the opportunity to talk about this important topic.

Dr. Lisa Hassler: 24:43

Great. Thank you so much. If you have a story about what's working in your schools that you'd like to share. You can email me at drlisarichardsonhassler@gmail. com or visit my website at www. drlisarhassler. com and send me a message. If you like this podcast, subscribe and tell a friend. The more people that know, the bigger impact it will have. And if you find value to the content in this podcast, consider becoming a supporter by clicking on the supporter link in the show notes. It is the mission of this podcast to shine light on the good in education so that it spreads, affecting positive change. So let's keep working together to find solutions that focus on our children's success.