Join Daniel Rivera as he sheds light on the rapidly accelerating world of Artificial Intelligence, better known as AI. Can AI help educators in their work? Absolutely! Is AI going to replace good teachers? Absolutely not! Daniel's observations, learning, and experiences combine to provide clear explanations about this technology, helpful analogies to provide context for understanding, and practical suggestions that can be put into practice today. And a great teacher is mentioned as well! Another episode, featuring more conversation with Daniel, will be released soon.
3:25 – AI has the potential to reshape every industry, job, and life
4:30 – the shift from narrow, programmatic control to individual user control
7:00 – a democratized internet with language comprehension
8:00 – the move from menus to genies
10:00 – you can ask AI virtually anything
10:40 – offsetting the unfortunate reality of the suppression of curiosity
12:00 – Hallucinations, mistakes, guessing, and precision
14:15 – AI iterations are getting better
15:20 – Bias in AI
16:10 – Data matters
16:20 – Notebook LM does not access the web
16:45 – User instructions, or a company’s training, can bias AI
18:15 – Engagement as evidence of curiosity
19:00 – The spectrum of apathy to curiosity
21:55 – Responsible use of AI for formative assessment support
22:45 – AI used with curiosity is a tremendous mentor, coach, and peer
22:55 – AI used with apathy is a threat to your brain . . . it might “make you dumb”
24:40 – You must be aware that using AI in a bad way is very detrimental – AI is not going to tell you this itself
26:45 – How much work do students do that generates pride and satisfaction that leads them to keep that work?
28:40 – All hope abandon, all ye who enter here
29:00 – Great teacher Ms. Smith, Statesboro High School
29:50 – We need artists in the classroom – creative, passionate, flexible
31:10 – a mistake to create a fully AI school
31:30 – Is AI a threat to teachers?
31:50 – AI will never replace good teachers
32:10 – Good teachers are curators of learning experiences
32:55 – Decisions about assessment, pacing guides, etc., are not always meant to benefit students
33:10 – What is the value of consistent instruction?
35:50 – Recruitment problem or retention problem?
36:30 – Low-hanging fruit - break down a lesson into manageable parts
37:00 – Low-hanging fruit - be specific with the target audience for what is to be learned
38:00 – Make this old lesson better, I need resources that cost less than one dollar per student
38:45 – Ask AI to use witty banter, match student interest
39:00 – AI will use analogies, metaphors, and more
39:15 – Teachers still do all of the final curation and make all of the decisions
39:20 – Create songs with Suno, an AI tool
40:20 – “Can AI…?” Assume that the answer is “Yes.”
42:55 – You have a genie. Dream big.
43:10 – Tell AI “how to act”
43:40 – Another AI episode with Daniel Rivera is forthcoming
Host (00:11):
What matters most in learning? The challenge, the thrill, the benefits, interacting with other people, or something else entirely? What is the connection between leading and learning? Does change drive learning or does learning drive change? What's more important, teaching or learning? Is everyone a leader, a learner, a teacher? Want answers? Listen in as we address these intriguing issues through commentary and with guests who share their thinking and tell us their stories. Lead, learn, change.
Daniel (00:55):
Some people mistake AI or see it as a threat to teachers. I think that there's some schools that are moving towards a fully AI-driven school. I think that that is a mistake and a grave mistake at that. You think our kids reacted poorly to being sent home during COVID without any human interaction. Really? Just wait. How do I relate the parts of a cell in a way that my kids who love Minecraft will understand? How do I explain the whole Declaration of Independence in a way that's meaningful for the kids with relationships, pop culture and things that teens are into? You have a genie. You can ask it anything. You can dream big, and it's there at all times. You just have to tell it how you want it to act. Pretend to be my student. What would you complain about with this lesson? There's so much we could do with this if we just ask.
David (01:56):
Today's guest on Lead. Learn. Change. is Daniel Rivera. Daniel, thanks for taking your time to speak with me today.
Daniel (02:02):
Yeah, thank you for having me.
David (02:04):
Daniel's an expert on the use of AI (artificial intelligence). He's been working with teachers for about two decades now, ensuring the educators are aware of current innovations and are equipped to maximize resources for the benefit of learners. Daniel's a tech director with the First District RESA – that stands for Regional Education Service Agency – and that serves 17 districts in southeast Georgia with Screven, Jeff Davis, Glen and Savannah-Chatham as the corners of that region. Daniel's a great thinker and he'll be sharing some of his knowledge with us today. So are you ready to get started?
Daniel (02:40):
Yes, absolutely.
David (02:41):
All right. Daniel, there's probably nobody listening that hasn't heard of AI and knows what it stands for. At the same time, there are likely some misconceptions about what AI is. Is using AI just the next latest, greatest, “one more thing” that teachers are being asked to do ,or are there good reasons to consider using AI with our students and in other areas of our work?
Daniel (03:06):
Well, I would say on the surface it may appear that it is just one more thing. It may appear that it's just another passing fad or more silicon snake oil. But no, it's far from it. It's potentially the biggest instructional technology that I've ever seen in my life and has potential to reshape every industry and every job and every life. It's a big deal.
David (03:33):
So it can be a huge addition, a big help to the array of resources that teachers already have access to, and they can use this for lesson and unit design. And one of the biggest things I've heard you talk about is the value it has for customization and differentiation. It's really a virtually limitless pool of ideas that line up with the student's interest and their current capabilities. And so that could be a great starting point for a teacher that wants to test the AI waters. But it's so much more than a quiz bank creator. When I spoke with you earlier, you shared some information that made it obvious that it's the tip of the iceberg thing that most of us are familiar with. Can you run with that idea of infinite examples?
Daniel (04:16):
Yeah. I think the biggest thing we have to kind of understand with modern AI is that it's not your father's AI, your mother's AI, it's not even yesterday's AI. The new shift to these large language models, what we call LLM, really opens the door to how we interact with this technology. In the past, we were so limited in what we could do with artificial intelligence because it was in devices or systems that we didn't really have any programmatic control over. What I mean by that is if it's in Siri for example, or Google Home or in this sales software, it's programmed to do a very specific thing. It's artificial intelligence within that and you only have a certain amount of commands and so on. So a way to kind of explain that to people who are new to this is that’s it’s kind of like the internet before 2000.
(05:16):
If you are old enough to remember that, it used to be that websites were kind of read-only and you had to either know HTML and have access to the server and the files or you had to be wealthy enough to pay someone to do that. If you had a website, you either knew someone or paid someone or were someone who was proficient in HTML who could do that. It's almost like the printing press. The Scribes Guild were the masters of information until then. And so in 2000, the internet changed forever, roughly around 2000, when XML came out and it was just a different way of writing webpages, but what it did is it separated the content from the form. Form and content were different now. So whenever you did a header, you didn't have to tag it as “header one” or not even that. You didn't tag it as bold and 20 point font and everything else.
(06:12):
You now could just tag it as “header one” and then in a separate document it would describe what “header one” is and the average user didn't even have to put in “header one.” The average user could then just fill in a box like a blog software or Facebook or not even then, MySpace, and it would then take that information and put it into the website and do all the programming for you. Well, what that did is it allowed you and I and everybody else who didn't know HTML to suddenly contribute to the web. So you go from having a 1% (if that) part of the population being able to control and contribute to the web to suddenly everyone and their mother literally could contribute to the web and that democratized the web. So with AI today, we now have included a second component to that, machine learning, in the form of language comprehension.
(07:16):
So AI now can understand and be programmed with plain English or 50 other languages in regular speak, using analogies, using turns of phrase, idioms and so on. But the point is you can now just talk to it and ask it almost anything and you're not driven by a finite set of choices anymore. So I've used an analogy of menus and genies. We live in a world of menus. We live in a world of limited choices. Even when you open up your phone, you have a certain number of icons you can tap on when you're on your computer, you have icons and menus that you can select, but that's it. Rarely can you just do anything with a technology. Now, if I were to say that you're used to these menus, let's say you go to your favorite restaurant and you have a certain dish that you just love and one day they came to you and said, “We have a genie in the back that can cook anything you desire, what would you like?”
(08:22):
A lot of people would be faced with decision paralysis and they would just kind of say, “Yeah, give me the number seven combo, the Swiss burger. I love it.” And you would just fall back to what you know. Some people might even think a little outside the box and say, “Oh, I've always wanted to try this one dish from China,” or something like that. Very few people would really get outside the box and say, “I want to steak made from the most succulent beast three galaxies over and it's going to come with a starch and a sauce that I know you will know I love because you're a genie and it's never been made in the history of the universe, but you're a genie. You can do that.” So we have to start thinking in terms of genies and not in terms of menus. And we have never done that with technology before.
(09:11):
All technology has been very limited. And so it's kind of like suddenly a couple things happened here. One, anyone can talk to AI now, and that completely opened up 99.9% of the population to AI that didn't have access before. You can program it and tell how you want to act and there's no menu so you can start asking it for any number of things. And so the kind of discussions I'm having with ChatGPT now are kind of just “out there.” Some of them don't mean anything really to anyone else but me and some of them are actually quite good for my profession. So whether I want it to create a parody song . . . I recently saw on Facebook that someone said, “My daughter misunderstood M.C. Hammer as McHammer and thinking it was a Scottish singer. And so someone started writing a Scottish version of You Can't Touch This, and they did a couple of stanzas and I was like, “Hey, ChatGPT, continue this song.”
(10:13):
And then I had it generate images of a Scottish looking M.C. Hammer and everything else. That's just silly, right? But it's fun. Or whether or not I'm arguing with my wife about whether bar soap is as effective or good for you as liquid soap. (It is by the way) or anything else like that. Any time now I can ask questions that I would not have asked in my adulthood because I've had the curiosity beaten out of me in a sense. Most of us have, but the kind of questions you would ask as a kid and there's no one there to say, “not now,” “Ask again later,” “We don't have time for that,” ‘Baby, you’ve got to get to bed,” none of that. It's always there. At 3:00 AM I can ask questions and entertain ideas and bounce ideas around and even say check my thinking on this. So that's that part of that genie thing. We aren't thinking big enough in terms of what it can do and we're just talking chatbots right now. We're not even going into all the other AI stuff. So just chatbots,
David (11:17):
Just the notion of genie versus menu is huge. That's a huge mindset shift because it really is a “the way we've always done it” kind of thing.
Daniel (11:28):
Yes.
David (11:28):
Now with technology, and I know that word means different things to different people, it really just means a tool to get something else done. But with this digital technology, those options have opened way up. It's interesting how we end up pulling in terminology from what we already know and what we're already familiar with to describe a new phenomenon or things that happen inside the latest iteration of a technology. You said that this is not your mom's, your dad's ai, you said it's not even yesterday's ai, so it's moving so fast that you can legitimately say that. Talk about hallucinations and the bias or shortcomings of AI detectors and what those things are
Daniel (12:09):
As far as the hallucinations go, it's interesting that we have a different name for them than we do with humans. I mean AI is just as prone to mistakes as humans are, and we're thinking it's a computer so it can't get it wrong, but if you think of it more like a human, and it's not really a human, but more like a human, then you understand that humans make mistakes and humans misremember and humans have bias. So let's address the hallucinations first. The hallucinations are really the imprecise memory of the AI. Now that comes from a lot of different possible places. One, it could be the data set. A lot of the AI there is trained to answer your question and be helpful and it doesn't have precise knowledge of some of the stuff it was trained on, similar to people. Let's say you have a favorite book, mine appens to be Lord of the Rings, and you were asked, “What exactly did Gandalf say at the Council of Elron concerning his time imprisoned?” Well, I don't remember exactly what he said.
(13:17):
I can't quote exactly, but I think he mentioned this, but I don't know. A human might say, “I don't really know,” and all that, and AI may just go ahead and tell you an answer and give it its best shot. It's been trained on Lord of the Rings but it's been trained in the same way a person has been. It can't cite page and verse unless it's absurdly relevant or absurdly well known such as Shakespearean quotes like “To be or not to be. That is the question.” That it can probably cite pretty reliably, but other things it can't. And that's something to keep in mind when you say, “I want the exact standard for seventh grade on erosion.” It may not be able to give you exactly that, and it may try anyway unless you even tell it, “It's okay if you don't know exactly.
(14:07):
I'd rather you be precise than helpful.” You're effectively telling it. You're still going to get the carrot and you're still going to get rewarded for making me happy because what makes me more happy than an answer is the precise answer. So sometimes it's how you prompt it. Other times, depending on what AI you're using like ChatGPT or Perplexity, they can access the internet and so can, most of 'em by the way, at least on the paid plans or on the paid features, you get a few times a day for free. So if you have web access, you can tell it, “I want you to search the web and cite your sources,” in which case your hallucinations are way, way, way less. Having said that, also GPT-4, the paid version is way smarter than GPT-3 0.5. You might say that they're the equivalent to a smart elementary student at 3.5 and a smart high school student at four, and version five is in the works and it's probably going to be as smart as a fifth-year PhD.
(15:04):
But the point is there are a lot of these variables that are affecting that hallucination thing, and so getting it to site its sources, getting it to search the web, asking it more general questions rather than precise questions, giving in and out. All those affect that. As far as the bias goes, it has the same bias as humans. In fact, probably less because it has less emotional baggage associated with its bias. Almost everybody has some sort of emotional response to things and has some sort of opinion. And the stronger we are affected emotionally, the more likely we are to be inaccurate or biased because of that emotion. You have to look no further than current politics to see that. So AI is also biased in its dataset. For example, if it's been trained primarily on Western media, then it's going to be inherently biased towards that information.
(16:00):
If it's been trained on university media, then it's going to have a slight or great, depending on the data, political bias in one way or the other. So the data matters. And that's another good thing about some of these paid services where you can upload or attach your own files or dataset or something like Notebook LM, which does not access the web, does not use its own training, looks only at the data you provided and will even tell you that's not in the dataset that you provided to me. Sometimes you use a certain tool that you know is centered around the data you give it, and then it's got the bias inherent in the data you give it not that bias inherent in the data open AI gave it or something like that. The other source of bias is overt instructions given by the person who's using the AI or the company that is training the ai.
(16:51):
The most recent example of that is with Google and Gemini, they had a pretty embarrassing fiasco with the bias in Gemini and they had to come out and publicly apologize for it, and they've been making some corrections since then, and I think they've actually done a pretty good job of bringing it back to center, but it was embarrassing for the company just how biased it was. But that was a well-meaning attempt to try to correct the bias in the data and they wound up overcorrecting with the bias in the training. But that's the case of a lot of times where humans try to intervene in a very complex system like habitats or ecosystems and they overcorrect.
David (17:32):
Being aware of the potential problems and knowing you need to be clear and specific and unbiased in your prompts and what you're asking AI to do and being aware of the limitations of the sources it can pull from will be very, very helpful when you're designing lessons or units or brainstorming ideas and that sort of thing. I think over-reliance on it, once we see what it can do is also problematic because we might erroneously make the assumption that if we use AI, that students are going to be engaged in the work. And when you and I spoke earlier, you talked about engagement not really being the goal, but it's almost a lagging indicator per se, that it's just evidence that the real goal has been met and the true goal is the student being curious about what they are going to learn or something that they hadn't thought about before. And you said something that was really, really interesting during a conversation earlier, which was curiosity is the mortar that holds together the bricks of learning. You talked about the continuum between apathy and curiosity as opposite ends of that spectrum. So share some of the thinking about curiosity and how AI feeds into that. We want to create the conditions that lead them to be curious so that they're engaged on their own.
Daniel (19:06):
Yeah, yeah. I love your term lagging indicator. Engagement is a symptom of an emotion or a mental state, and that is of curiosity, and we often want the end result. So we demand the end result when we really want the conditions that produce that result. Another comparison I say is you go to your spouse and you say, “You don't tell me you love me very much anymore or you don't buy me flowers anymore.” And then they say, “Oh, well here you go, I love you,” and they hand you some flowers. Are you happy? And then they walk off.
(19:49):
It doesn't mean much when it's just done. You don't want that behavior. You want what's behind that behavior. What we're really wanting is not the engagement we really want the curiosity. If we just look at the engagement and we just assess for that, that can be faked. That can be faked by the students, it can be faked by the teacher, it can be, “Everyone be on your best behavior today because I'm being observed,” and they are and they look great and they learn no more than they did yesterday. We have to be a little careful with that. So curiosity is a big one. Now, the whole spectrum here, curiosity is one end of a spectrum. It's a component on a spectrum. It's like hot and cold. There's curiosity and apathy and apathy is the absence of it, of course. You just don't care. You're not interested, you don't want to know.
(20:38):
You might say it's incurious in a sense, but apathy is a little bit more, I think incuriosity, might fester into apathy eventually. So apathy is the mind killer (to use a Dune reference). It's not fear, it's apathy. And the moment you become apathetic, you cease to learn, you cease to care, you cease to grow mentally, and AI thrives on both. If you're fueled with apathy, you will not want to do the assignment, you will not want to write the paper, you will not want to generate the learning goals and the report you have to do on Monday for your business. And you just want the pain to go away and you just want to get back to your show or go hang with your family or whatever else. And you'll seek whatever salve that you can find to make that misery go away. And AI is there to swoop in and do your work for you.
(21:37):
And sometimes that's a good thing because you buy time and you really need your time and you're not really growing in that assignment. Let's say this is the fifth time you've had to give feedback to students or the 50th time you're giving feedback to students. That kind of formative assessment, does it have to come from you? Can it come from the machine? Yes, but there's a risk in that if you're not reading your students' work, then you don't know them. So in that case, you could be apathetic and let it grade entirely all 50 students' papers and give feedback or you can use that as a co-teacher and you read their papers and you let it give initial feedback and you decide if you agree or not. That's a lot less mental load than doing it all yourself and then having to wordsmith for the 50th time something about splices.
(22:28):
You just have to agree or not or make a slight change and if they’re good, but you still did half the work. You met the AI halfway. And then the other way is if you're learning, if you're a new teacher and you need to learn how to give proper formative feedback, then you may get the feedback and have the AI analyze your feedback and tell you if it's good or not, or you may do it for the first few of them and then say, “I think I got this,” and then you move on. So AI, if you're using it with curiosity and you're using it with a growth mindset, it's the best mentor and tutor ever made by humans in the history of the world. And if you're approaching it with apathy, it's the worst chronic threat to your brain I can think of because it will make you dumb over time. You'll become dependent on it because you will not have worked out your brain. It's like the spotter at the gym or your physical trainer at the gym doing all the workout for you while you sit and eat Cheetos on the side.
(23:29):
You're not just skipping leg day, you're skipping brain day, you're skipping everything. And I worry that first year teachers are going to fall into this problem where it's writing other lessons for them. 30 year veterans having it write a couple lessons and they look at it and go, “This is trash. Do it again.” That's very different than first year teacher having to write a lesson and they don't read it or internalize or have it teach them, and they just have it churn out lessons for them for the next three years. Those are formative moments in their educational growth. And then you take a sixth grade student who uses AI to do all the thinking. By the time he hits 10th grade, he's still a sixth grade mentally and now he's dependent on it like some sort of drug just to get to the same point he should be at because he should have grown there.
(24:16):
So yeah, that's the apathy, curiosity spectrum and why I think that is the most important emotion you can have when you're using AI. And the biggest threat we have long-term, the real problem I'm seeing is that it doesn't feel bad to use AI in the bad way. It's kind of like sugar. It tastes great, your body craves it. Every kid in America or in anywhere loves sugar and your body will not tell you that it is too much or whatever. You might get sugared out if you eat an excessive amount. Let's say you overdose on sugar every day for 10 years, you're going to have diabetes, but your body doesn't really tell you that it's because of that sugar you've been eating. You don't make that connection. The only way you make that connection is through education. So your body will lie to you and tell you it's great and your brain's going to lie to you and tell you the AI doing all the work for you is great. The real danger is chronic, and curiosity is the antidote to that. The problem I'm seeing is that we're at stage three cancer with apathy in education. We're almost terminal. And a lot of the implementation of that of AI is going to be driven by apathy, and we got to get ahead of it. So I'm talking to every teacher I know and say, “We're going to have to change how we teach because our teaching is breeding boredom and apathy.”
David (25:44):
You're saying that also out of personal experience. The listeners might be interested to know that you didn't exactly love school as a student for period of time. And I think that you spent time in the woods, if I remember, with a good book and you skipped school pretty frequently. And now here you are in a career immersed in a career where you work with teachers nearly every day. So when it's used well, how can AI reach somebody like you were in that era?
Daniel (26:13):
Oh man, yeah. If I had AI, I would be so much smarter than I am now today. I was in the gifted program. I was no mental slouch. I was a pretty bright kid and I love learning. I just hate being taught. And I think that that's the thing. I really detested school because so much school was busy work in my opinion. And oh, you couldn't say that because you'd get in real trouble if you did. Teachers didn't like you to say that, and I get it. They didn't want their work to be diminished or demeaned like that, but a lot of our work that the students do in school is destined for the digital trash can. I mean, really think about it, how much work do they keep after they graduate and that they're proud of? I've got a shoebox, roughly a cardboard box in my closet, deep in my closet with papers that I wrote for the Quest program.
(27:08):
These were the independent studies I did in Quest, going back to third grade handwritten yellowing paper reports I did on reptiles, Norse mythology, optical illusions, and I keep them as keepsakes. When we were in the Quest program – that was the gifted pullout program once a week – we didn't address standards like classroom standards. We had our own Quest standards. They centered around things like creativity and imagination and elaboration and synthesis, all that. And we were given independent studies of anything we wanted to pursue. And then we had to generate essential questions and we had to take notes to answer those questions with at least so many bullet points. But there was a thing we were passionate about. I was really into Norse mythology, so I got to delve deep into that. And in the process of that, I learned how to write outlines. I learned how to do rough drafts.
(28:05):
I learned how to do all this stuff. That was the price of admission for my curiosity and my curiosity fueled that the entire time. And then we had to present it and we had to know it. We had to present it and talk about it and have visual aids and everything. And we were proud of the work we did so much that even now, at 48, I still have these things in my closet. I can't say that for a lot of the work I did in school. I do remember the leaf collection. I do remember band. I do remember art ,and an occasional paper. I remember going into the first day of senior year language arts, Ms. Smith, I can't remember her first name, Statesboro High School, Ms. Smith, and I walked in and on the board were the words “all hope abandon ye who enter here.”
(28:55):
And we were like, “What is this?!” And it was so shocking and so terrifying and so whatever. And then she came in and said, “I want you to write about what that makes you think of.” And then she later told us that that was written on the gates of hell in Dante's Inferno and all this other stuff, and dismissed this feeling that she was some terrible person, that she was just like, “I wanted to get a reaction out of you. I wanted you to think, I want you to write. I want you to express yourself. “And every day she came in with a different quote. “Resolve to live each day as if it were the last day of my life.” That was another quote I remember seeing and writing about. And she had a way of just encouraging us to express ourselves through word. So everybody approaches it differently, every teacher approaches it differently.
(29:38):
And there are some good teachers I remember from there and even in college, but there were a lot of other teachers that meant well, but were not either allowed to or were not inherently artists and we need artists in the classroom because artists will inspire and pique the curiosity of these students more. And I don't mean artists in the traditional way. I mean artists in the sense of a good teacher is creative and knows their craft, but is allowed to express it. They have a passion. They were willing to modify their instruction or whatever based on the needs and the personalities of the students that they're working with.
David (30:28):
Think about two coexisting principles and talk to us about these for a minute. The value of a great teacher cannot be overemphasized. And two, there are some things that great teachers traditionally do that AI can truly help with and help greatly. For example, question generation or providing feedback, which you mentioned earlier. Just nudging that learning forward a little bit with just in time reminders or maybe getting a start on developing a rubric and differentiation and even tutoring. We'll talk about conmigo later for example, but this is a jumping off point for an opportunity to address the paramount value of great teachers and the legitimate value of AI and how they can work together.
Daniel (31:13):
I think that there's some schools that are moving towards a fully AI driven school. I think that that is a mistake and a grave mistake at that. You think our kids reacted poorly to being sent home during COVID without any human interaction. Really? Just wait. So I say that because I think that some people mistake AI or see it as a threat to teachers. It can be a threat to burned out or mediocre or checked out teachers, maybe apathetic teachers, it can be, or those who are so hemmed in by what they're allowed to do in their classroom, it may replace them because it can, but it's not going to replace those good teachers. Good teachers are able to let the AI drive the car but then take over at any time. And when they let the AI drive the car, they direct it to where they want it to go.
(32:10):
It's less about the driving and more about where you're going. And so good teachers are curators of learning experiences and they either contribute to them as a live actor, maybe that's your thing. Maybe you are just really dynamic and you are Mr. Keating from Dead Poet Society, you've got the energy or maybe you're a little less on that and you carefully construct almost Rube Goldberg like scenarios where this learning leads to this learning leads to this, leads to this thing. Every teacher's going to have their own way of handling that. And if they're really good, it's effective and the kids love it. And it'll be different in every classroom if they're allowed to do that. Unfortunately, and it's going to get me in trouble . . . but unfortunately, a lot of those decisions we make in terms of common assessments or pacing guides or so on, are not meant to truly benefit the students.
(33:08):
They're meant to make the administration and reliability the star, make it easier to administer, make it easier to assess, make it easier to get those numbers and make it predictable. Now there's this discussion that, oh, well, the students are better if they have a consistent instruction. Well, how often do the kids change classrooms? And if they do a lot, then why are they changing classrooms so much? And even if they do change classrooms and they have a very different teacher from one classroom to the other, and they're at a different part in the year, okay, they miss a couple of questions on the standardized assessment, but maybe it's good for them to learn how to adapt because every teacher is going to be different in the next year and in college. And the price we pay for that is a paint by number type instruction.
(33:55):
And you want to watch people die inside, give creative dynamic artists paint by number. You're not going to get catastrophic art, but you're not going to get good art. And what you will do is replace their passion with apathy or resentment, and some of them will lead the profession within the first few years and other ones will get burned out, and that apathy is contagious. So the students will become apathetic and AI is right there to make everything go away. So right now, more than any time we need artists in the classroom, we need teachers who are able to teach the way they feel is important and still address every standard. You give every teacher the same standard and they're going to teach it very differently. This high school teacher's lesson on our unit on the hero's journey, the hero's story, or the plot structure is going to be different than this
(34:51):
teacher's is going to be different than that teacher’s. And if we're really doing standards-based instruction, it's really more about those skills and not about content. Most of the standards now, especially English and all, they don't say you have to teach the Iliad. They say you have to teach hero’s journey or epics. But I see again and again, due to trying to get some consistency and it's all well meaning, all the teachers in the grade team teach the same thing and do the same assessment even though they may very much disagree in how they approach it. I'm looking to see what we can do for compromise there because I understand the value of the common assessments and the pacing guides, but I also think that those really do kill creativity and they kill the artistry if we're not careful. We need to provide a lot more flexibility within that system, and each system will have to find its own answer. But I can tell you this, if we keep doing it like we're doing, we're going to keep losing people in the profession. It's not a recruitment problem, it's a retention problem, and we're going to keep making school less and less dynamic, less and less creative, less and less passionate and more filled with apathy.
David (36:04):
Think about some of the places where you've seen where you would say, “this is a success story.” Not because AI is being leveraged in the most efficient, effective way possible, but just where these great teachers are using AI, in I guess I would call it a masterful way, so that students are eager to learn and learn some more, and then maybe share. So if you've got an example or two from elementary, middle, or high or all three.
Daniel (36:30):
Yeah, sometimes it's a matter of just breaking it down and making it more relevant for students and easier for them to understand. So that's their low hanging fruit. That's an easy way to use AI without having to really change your instruction intensively. Sometimes it's a matter of saying, “Look, here's a lesson plan. I'm attaching it. Take a look at it.” And you're talking to the ai, right? You're saying, “Take a look at it and I need a better opener. I need something really dynamic to help snag the kids. By the way, they're really into this or this, right?” I saw someone trying to explain MTSS to first year teachers or new teachers and they happened to be coaches. So I said, “Oh, yeah, let's try that.” I tried it with AI. I said, “Explain MTSS multi-tiered support systems” or something like that. I'm not an expert in that.
(37:09):
“Explain that to me as if I were a first year teacher and coach. I don't have much time, so be brief, use analogies to sports, so I understand,” and it goes and does this really short three bullet thing. Talking about tier one is the game plan. The whole team gets trained on that. Your standard practice, everyone gets the same high quality instruction in the classroom. It's your general teaching strategies that help most students succeed. Tier two or extra drills, small group work. Some players need extra practice to master certain skills like shooting or passing and so on. And similarly in MTSS, small groups of students get targeted help in specific ways. Tier three, personal coaching, that's intensive individual help. Few of those players, very few need that one-on-one coaching to address significant challenges and so on. I put that out there and the MTSS experts said, “That may be one of the simplest, most concise descriptions of MTSS I've read.”
(37:57):
I was like, great, great. So you can use that with your adult learners, right? Sometimes it's a matter of having it broken down for this student. So these are low hanging fruits, right? This is how you can implement AI tomorrow. “Here's an old lesson. Make it better.” “Give me ideas for this.” “I need a hands-on project to help the students learn this concept that won't cost me an arm and a leg at Walmart. Go ahead and break down the estimated prices based on current prices nationally or go to walmart.com and look up the prices for all the materials. I need it to be less than a dollar per student.” Whatever. I mean, why not? Or, “How do I relate the parts of a cell in a way that my kids who love Minecraft will understand? How do I explain the whole declaration of independence in a way that's meaningful for the kids with relationships, pop culture and things that teens are into?”
(38:43):
And sure enough, it spits out, “Oh, well, yeah, write it as a breakup letter. It's supposed to be a give take relationship, and lately you've just been takin’,” and it starts writing that. And it's funny. “Use witty banter, use analogies, be funny,” and it does it, and you just say, okay, that was a whole lot of energy I didn't have to do last night in preparation for today, but I'm selecting what answers I choose to use in my classroom. I can throw all this away or I can keep some of it. And that selection, that curation is an important part of this. And so I'm seeing teachers do things like that. I'm seeing teachers use something like suno, which is songwriting AI. It generates lyrics and everything, full blown song, music, everything. And these are bangers too, and it creates a song about anything. I saw a teacher make a song, a death metal song, about linear equation.
(39:38):
It was the funniest thing, but yeah, it was great. I made a song about the solar system that had kind of a bubblegum pop sound to it. We've made songs about anything, and the kids absolutely get pulled in by that. I've seen teachers use AI to generate images for their presentations. That's another low hanging fruit. Simple thing. But now they've got a cool image of historical people or whatever else, and they don't have to then go searching for two hours to find the right stuff for their presentations. Or do an epic rap battle between Lincoln and Douglas, right? There was an incredible debate. Upload the debate, turn it into a rap battle. Go. Why not? I've never tried that, but I assume it's going to work. Why? Because the answer is, “Yes” 90% of the time. Can AI do? Yes, just assume the answer is yes.
David (40:25):
You're describing the right mix of the concept you mentioned earlier about menus and genie. So the menu is “Here are standards, here are concepts or content that must be addressed that students need to understand and be able to demonstrate some competence in.” And once you say, this is what it's about, then whatever it takes to appeal to different motives and interests that students bring to the table about those things is wide open and it's going to, each iteration is going to serve as a springboard for your next ask. So this notion of the human element in reviewing what AI generates or spits out and refining the prompt and then refining it again, and obviously verifying sources and that sort of thing (if it's allegedly factual content), asking AI back, “What else do I need to provide to make this prompt better or analyze my assessment practices?” Like you mentioned earlier, the possibilities are endless, but we have to remember going back to this great teacher notion that it's the good teachers who have been doing this on their own with all this effort, and now they can have someone, not someone, some tool come alongside them and support their efforts and it doesn't s squash their creativity. It actually expands it with, so you get to do more with less effort.
Daniel (41:59):
Well, it is an amplifier, just like all tech. If you bring apathy, it's going to amplify. If you bring curiosity, it's going to amplify and how much you bring, it's almost like an n to the nth. You bring one unit of this to the table, it's going to give you one, you bring two, it's two squared, and so you get four, you bring three, it's three to the third. Now you get 27, right? And you bring 10, it's 10 to the 10th power. It's crazy what you put into this with AI comes back in such huge ways, and so you have to bring some creativity and passion to it. What I'm seeing, just like we see with all technology when it's first use is, “I'm burned out. I'm tired. How do I use this to make the pain go away? And how do I use this to do everything I used to do slightly better?”
(42:50):
We need to go beyond that, though. It takes time to get beyond that, but we need to remind teachers, you have a genie. You can ask it anything. You can dream big and it's there at all times. As a co-teacher, as a peer, as a parapro, as a coach, as a guide, you just have to tell it how you want it to act. As a student, “Pretend to be my student. What would you complain about with this lesson?” I don't know. Have you ever tried that? Let's try it. Let's see what it does. We could be using AI for teacher prep. We could use AI for behavioral interventions. There's so much we could do with this if we just ask.
David (43:30):
This is so interesting, and we could keep on talking, I'm sure. So are you willing to share more insights in another episode? I'll make sure that happens. What do you say about that?
Daniel (43:39):
Absolutely. Yeah.
David (43:40):
Excellent. You've already shared some great ideas about getting started with AI, and the show notes will include summary points so that listeners can go find the topics we've discussed. Thank you, Daniel, and have a great day.
Speaker 4 (43:53):
Thank you. Thanks for listening today. Find the Lead. Learn. Change. podcast on your search engine, iTunes or other listening app. Leave a rating, write a review, subscribe and share with others. In the meantime, go lead. Go learn. Go make a change. Go!