Edtech Insiders
Edtech Insiders
Google’s Vision of a Personal Tutor for Every Student: Dave Messer on Guided Learning
Dave Messer is a Product Manager on Google's Learning & Education team, leads product for emerging learning initiatives, including Learning Labs and Gemini. A former teacher with masters degrees in software engineering and education, he works with experts and the education community to build products that target real-world challenges for educators and students.
💡 5 Things You’ll Learn in This Episode:
- Google’s vision for AI in education
- How multimodal learning makes concepts stick
- Why student feedback shaped Guided Learning’s design
- The role of learning science in AI models
- How Google balances personalization, privacy, and classroom impact
✨ Episode Highlights:
[00:01:24] Dave Messer’s journey from classroom teacher to Google PM
[00:02:31] Google’s vision: a personal tutor for every student, a TA for every teacher
[00:05:27] Multimodal learning in Guided Learning—diagrams, visuals, and YouTube videos
[00:14:16] Inside LearnLM, Google’s AI model infused with pedagogy
[00:19:33] Psychological safety and student agency in AI-driven learning
[00:23:15] Student feedback that inspired Guided Learning’s design
[00:27:13] How learning science principles like active learning and metacognition guide product design
[00:35:18] Personalization, data privacy, and the classroom of the future
[00:38:32] Educator concerns, student creativity, and the responsible use of AI
😎 Stay updated with Edtech Insiders!
- Follow our Podcast on:
- Sign up for the Edtech Insiders newsletter.
- Follow Edtech Insiders on LinkedIn!
🎉 Presenting Sponsor/s:
Innovation in preK to gray learning is powered by exceptional people. For over 15 years, EdTech companies of all sizes and stages have trusted HireEducation to find the talent that drives impact. When specific skills and experiences are mission-critical, HireEducation is a partner that delivers. Offering permanent, fractional, and executive recruitment, HireEducation knows the go-to-market talent you need. Learn more at HireEdu.com.
Every year, K-12 districts and higher ed institutions spend over half a trillion dollars—but most sales teams miss the signals. Starbridge tracks early signs like board minutes, budget drafts, and strategic plans, then helps you turn them into personalized outreach—fast. Win the deal before it hits the RFP stage. That’s how top edtech teams stay ahead.
As a tech-first company, Tuck Advisors has developed a suite of proprietary tools to serve its clients better. Tuck was the first firm in the world to launch a custom GPT around M&A.
If you haven’t already, try our proprietary M&A Analyzer, which assesses fit between your company and a specific buyer.
To explore this free tool and the rest of our technology, visit tuckadvisors.com.
[00:00:00] Dave Messer: Gemini 2.5 Pro is actually, we did a scientific paper, a technical report on this, but actually it shows like among teachers, it's preferred as a model. It is the most pedagogically infused model of all the models, and so again, this is the work of the past few years. We're continuing to focus here to make it more helpful to teachers and students, but really those principles of.
Making our technology helpful for students and teachers for learning and teaching really goes like full stack. It's not just about the surface level and like what we could do in a product, but really how do you think in the full stack from a model through the product experience. To connect to users.
[00:00:44] Alex Sarlin: Welcome to EdTech Insiders, the top podcast covering the education technology industry from funding rounds to impact to AI developments across early childhood K 12 higher ed and work. You'll find it all here at EdTech Insiders.
[00:00:58] Ben Kornell: Remember to subscribe to the pod. Check out our newsletter and also our event calendar. And to go deeper, check out EdTech Insiders Plus where you can get premium content access to our WhatsApp channel, early access to events and back channel insights from Alex and Ben. Hope you enjoyed today's pod.
[00:01:24] Alex Sarlin: We have a very, very special episode of EdTech Insiders. This week we are speaking with Dave Messer. He's a product manager for Google's. Learning and education team, and he leads the product for emerging learning initiatives, including Learning Labs and Gemini. He's a former teacher with a master's degree in software engineering and education, and he works with experts and the education community to build products that target real world challenges for educators and students.
Dave Messer, welcome to EdTech Insiders.
Dave Messer: Thanks Alex. So excited to be here and excited to chat.
Alex Sarlin: I am so excited to chat with you as well. We got a chance to do a, a little short off the record chat a few weeks ago just on the cusp of the launch of Google's guided learning project, which is totally amazing and I'm really excited to be able to get down now and actually get into it and see how things are rolling out and just open up the doors and talk about some of the amazing things that Google is doing.
So you mentioned in your bio that you are somebody who is a teacher, you are a teaching background, you're now a product manager at Google. How do you and your team think about the role of technology in shaping the future of learning, and how do you bring that teacher perspective to bear?
[00:02:31] Dave Messer: So when I was teaching, I actually saw, like, so much potential for technology.
I was incorporating it, using it, trying to, to use it to help my students learn, using it to help teach. And I was always driven by this curiosity of what else could we do? How could we use this in ways that are meaningful and help people learn, help personalize learning, differentiate learning? And that's been like the core question that I've been asking myself for years.
Like, how can we use technology to help people teach and learn? And that really led me to Google where what's amazing, like Google is no stranger to this too. The mission at Google has always been to organize the world's information, make it accessible to everyone and useful. And like the key thing there has been learning.
So whether you're searching to find something using YouTube to learn. Learning has been core in our DNA and like. Really the thing that we think about is how do we bring the best of technology with the best of what we know about what helps with learning with learning science, and pair those together to be as helpful as we can, both to students and teachers.
Our vision ultimately is like, can we help have a personal tutor for every student that can personalize learning and a TA for every teacher? So with AI, we feel like we have more opportunity to get closer to that vision than ever and be helpful to people. But yeah, I think it's really about that core question, like how can we use this in a way that's helpful to people and can help them both teach and.
[00:03:52] Alex Sarlin: Yeah, and I mean one of the really interesting things about how Google works in EdTech is that Chromebooks and now NotebookLM, but Gmail and YouTube have been absolutely core to how schools operate and how students operate and how teachers operate for many years now. You know, we talk on the podcast about how Google is sort of basically the biggest education company in the world, even though it's not the.
Absolute core mission of Google. You have this enormous community of teachers, an absolutely massive community of educators all over the country and the world that are available to give feedback, to help build products, to test products, to give ideas for new features. I'm curious how that impacts your work, you just having this enormous, really engaged community of educators.
[00:04:35] Dave Messer: Totally. I think it's a great question and like our work really is for people, right? So one of the core values at Google is to respect the user. And what that means is we listen, we talk, we try to understand what people need and follow that to help create things that are useful for them. From our Google Champions community who's amazing, they're like, they show us new ideas for how they're using the technology.
They're the ones who can, like, find the gap or hey, it doesn't quite do this thing and we need it to do it so that we can help students in this way. Those are the guiding light for us to understand what people need, how can we be more helpful? So those conversations, those relationships, whether it's students that we're working with, whether it's that community that's really like the heart of what we do.
It's for people and we wanna understand what works and then work with them to create products that can be helpful.
[00:05:27] Alex Sarlin: You have such an engaged cohort of teachers between the Google certified educators and people who are champions of various aspects of the Google community. You mentioned YouTube in passing there, but one other aspect of Google's learning strategy that I found really exciting is this leaning into multimodal content, voice activated learning translation through visuals.
Obviously YouTube videos and guided learning has a very specific. Philosophy on how it uses multimodal content, especially imagery in its tutoring. When you ask it a question and say, I wanna learn about something. It's really specific and thoughtful about how it uses multimodal content. I'd love to hear you talk about the philosophy behind that.
[00:06:05] Dave Messer: Yeah, so I think like thinking back to my days in the classroom, if I was just explaining things or just having people read things and all they saw was text, like it doesn't click. It doesn't always make the connection that you need. And so learning does work, takes a little bit of actual friction to like make those connections, think and understand things.
But what we're trying to do is like how do we minimize the unhelpful friction? How do we help you get these like faster time to your aha moment where it clicks and you understand it and really explain that. So how can we use different. Capabilities with technology that can help learners learn, understand things, explore things.
And I think the visuals and videos in guided learning are One quick example of that, right? Where to understand the parts of a cell. Like you could read it in a bulleted list, but it's not gonna click the way. If you see a diagram, you can ask questions about that diagram. So that's really like a bit of the philosophy.
How do we help make learning which is inherently hard, easier so that people can understand things go deeper, leveraging the parts of technology that we can use that are helpful in those moments?
[00:07:06] Alex Sarlin: And so if you ask Google guided learning about the parts of a cell, you say, what is the mitochondria? What are ribosomes?
I'm trying to remember my biology. Yeah. You know, and it realizes using its intelligent model, Hey, this is a great opportunity for visuals. How does it tackle that task?
[00:07:21] Dave Messer: So essentially what our work was, was to help the model, identify moments where a visual aid would be helpful and appropriate. It goes and looks for really high quality visuals to include and then brings those into the conversation in a way that can help you explain, just like if you're a teacher, you have a presentation, you wanna show the cell on the projector, or something like that.
It's that same moment. How do we help teach Gemini to be helpful when you have questions about the cell? I actually was like on my walk today, I was just curious. I was like, Hey, how do different animals breathe? And I was using guided learning to have this interactive conversation. Yeah. I started learning about how insects breathe and it blew my mind.
It was so cool. But the really helpful part was seeing these little diagrams to understand how they actually take in oxygen, how they breathe, like. Amazing. I actually like walked away. I understand it a little bit more now. I could probably go deeper, but those experiences and those multimodal connections really helped me understand something in a way that sticks.
[00:08:19] Alex Sarlin: Oh, absolutely. So that's the image, the sort of static image side. And then what about the video world? You obviously have Google has, you mentioned YouTube. It's basically the largest learning library in the world for video. Google has access to it, has complete understanding all the metadata of it. Do you surface YouTube videos within guided learning?
[00:08:37] Dave Messer: We do suggest YouTube videos as well as you're going through guided learning. Again, like we've really worked to make Gemini and guided learning make those decisions about what an appropriate moment might be for those videos, but really like when it finds the right moment to suggest, Hey, here's a great video that you can watch to get an overview of this or to dive deeper.
Those are the moments. It's really cool when you see a video, sometimes that is the most helpful way to understand something, like having an experience, seeing something in action. Even getting a walkthrough, like let's say you're stuck on a math problem, you actually wanna hear the steps step by step.
See it drawn, like, it's amazing. There's so much great content on YouTube. So if we can help bring those into a conversation to be helpful, it's great in some cases, like technology can't replicate that same thing like, you know, we were talking, I'm learning about erosion or the Grand Canyon, like how cool to actually see it and have this experience where you're like.
Get close to understanding something through as close to a real world experience as you can get from anywhere you are. So I think that's like the power of YouTube is incredible and just bringing those human voices into the conversation so you can learn from other people, I think is like a part of guided learning that is just trying to be helpful and help people make those connections.
[00:09:50] Alex Sarlin: Yeah, I would totally agree. And you know, one more question about the multimodal. This is like a passion of mine and I just find it so interesting, but I think you have really interesting insights into it as well. We're at this. Interesting moment, as you well know where, especially within Google, Google is going from.
Its really entrenched incredible expertise in algorithmic recommendations and web search and being able to catalog all the world's knowledge and content and surface it, including YouTube videos or diagrams or websites. But it's also now creating these incredibly powerful tools to be able to create.
Video and create imagery with nano banana and, and vo and flow and all the different incredible models and tools. And I know that you think a lot about this, like how do you balance this idea of, I'm asking to see a cell, when do you look up a really high quality diagram of a cell from a trusted source?
And when are you gonna sort of draw me a picture of a cell that looks like. You know, an anime ship or put me inside a shell and pretend that I'm traveling through it. You can do this kind of thing with ai. How do you balance
[00:10:50] Dave Messer: that and where do you see it going? Such a great question. Yeah. I'm like consistently blown away by the new multimodal capabilities, whether it's voice, whether it's image generation, video generation, it's really incredible.
There's like a couple ways to look at that. So I like, as a former teacher, a question I ask is like, okay, how could I use this technology as a way to help my students learn and teach? So actually like one exploration we've done is like, how could VO be. Used as a way to create engaging content. That one example was you could create inquiry science episodes where you have two characters that stumble upon some phenomenon and then it's actually the work of the class to go and explore it on their own's hands-on work.
But like these little moments that are kind of engaging and can create and like set up a scenario if it doesn't exist, you could search YouTube, see if that exists. But if it doesn't right, we actually create it now, which is pretty incredible. Yes. And find something that's like. Suited for your classroom, for your students, that can be engaging and helpful.
So I think like on that creator's side, I'm actually super excited to see what teachers do with it, what they create, and finding ways of it being helpful for learning. I think the same thing for students. It's like, I think your question about where would say VO or image gen fit when we already have videos and images?
It's a really good question. I don't know that we've answered it fully, but I will say like again. The core thing is what is helpful to people? What problems do they have, and based on the capabilities of technology. Is there something that we could leverage this technology for that would help people learn more?
So like if, for example, let's say they're like diagrams of the water cycle, we can only really find it at a level that's like high school, but you're really teaching it at a second grade level. I mean, I'm sure there's actually great content for that too, but as an example, if there wasn't something that was really suited to your students.
I think these media generation capabilities could be an amazing way for both students and for teachers to get that same learning value of content, the value of an image where it can communicate a lot that words can't. They can get that like right there and get that help in the moment that's suited to them and personalized in these like multimodal ways.
So I think we're still early on finding exactly where. That fit is, but there's so much opportunity and potential, I think excited to see again how educators and students use it to help themselves.
[00:13:12] Alex Sarlin: I totally agree and I, I've gotten the first chance to play with a video overview feature of NotebookLM, which it was announced mid-year at isti, but the idea that you can upload content and then have it turned into a, basically an explainer video, and I've been playing with that as well, and it's absolutely incredible.
So the functionality to do exactly the kind of thing you're saying. We'll go look on YouTube or we'll go look at our diagram content and find the exact right thing. And if it doesn't exist, we can make it, and we can make it, and we can adapt it and customize it and put you into it or change the style or change the age appropriateness.
It's an incredibly exciting world. Let's talk about the learning science aspect of it. You have a pedagogical background, and one of the things that's really been so interesting about Google's strategy and AI with learning. Is that it created this specific model for learning Google's learn LM model.
It's a model trained to understand education and to respond in an educationally sound way, and that is the model that undergirds a lot of the Gemini products that are learning based included guided learning. Can you tell us a little bit about Learn LM and how it shapes what you do in your learning suite?
[00:14:16] Dave Messer: So we announced Learn lm, I think it was last year, which is our family of models where really learning science is baked in from the get go. We assembled a team in, I think it was 2022, to start to explore that question of how AI could help with learning? And we found it very early on, like it needs some work.
It's not really out of the box ready for learning, ready for being helpful for both teachers and students. So. What our work has really been, how can we bake in the capabilities for amazing teaching and learning? Directly into our models and bring those across our products, across Google. So I think when you think about a lot of AI driven development, it is really about defining those capabilities that matter and working on those.
Think about your benchmarks, right? There's all these benchmarks about, Hey, this model can now pass a high school exam and it can pass a college exam. Our question has always been, well wait. How can we help people pass those exams? So a slight twist on that same question, but its ability to answer things.
Yes, it's super important, but your ability to develop an understanding of that thing is actually the thing that we've been working toward. So Gemini 2.5 Pro is actually, we did a scientific paper or technical report on this, but actually it shows like among teachers, it's preferred as a model. It is the most pedagogically infused model of all the models.
And so again, this is the work over the past few years, we're continuing to focus here. To make it more helpful to teachers and students, but really those principles of making our technology helpful for students and teachers for learning and teaching really goes like full stack. It's not just about the surface level and like what we could do in a product, but really how do we think in the full stack from, um, model through the product experience to connect the users.
[00:16:03] Alex Sarlin: Yeah. Full stack learning. That's a great metaphor and I love that comment about the benchmarks are whether the AI can pass an exam. Is it a PhD student? Is it a college graduate? Has to do on a math Olympiad, but in learning, that's not the question we wanna know. Right? Silly aside, but I've been thinking a little bit about like the metaphors for ai and it's like if you woke up one day and there was like a robot in the corner of your room.
What does this thing do? Like, what's the capital of the Philippines? And it can, Hey, it, it knows. Hey, then you ask it, Hey, could you tell me a recipe for something? It's like, oh, it knows. It knows everything. And not only does it know everything, it could pass your, you know, AP bio test. It could pass the bar.
But when you say teach me something, it's like, oh, I'd love to teach you. And then it just outta the box. And I'm not talking about Geminis specifically, but outta the box. These LLMs don't act like teachers, right? They act like butlers. They act like they wanna do it for you. And there's a real switch in mentality to turn it from a servant to a teacher, or tutor that is not how teachers operate when they work with students.
And maybe there's a launchpad for the question about guided learning. When you are thinking about guided learning and you're infusing, learn LM models into it. What is the sort of personality or almost like the traits of guided learning as a tutor or as a teacher that you think about to make it a really effective teacher and not just somebody who's trying to help at all costs?
[00:17:19] Dave Messer: Yeah, that's a great question and you captivated my imagination with your robot examples, like the thing I've been picturing too, along those lines. I kind of picture AI is almost like picture this infinite library and like there's all this information that's great. And you could ask the library and, Hey, what's the answer to this?
They go look it up and just give you the information. But I think what guided learning really is, and what we're working for isn't, isn't that retrieval, it's just not about information. It's about. How can you actually have a guide that engages you, helps you get motivated, helps you find those pathways through the information and what's out there, both in those books and in the world, so that you can start to understand things and go deeper.
I think that's really the key piece. Specifically, like talking about guided learning, what helped inform the environment or the like, the design of it. We worked with students and students are incredible. I know there's a lot of conversation about, hey, AI is being used for cheating. When we talk to students, they're actually using it to personalize their own learning.
It's amazing. They're so creative. They wanna learn. They're using the tools that they have available to them to learn in ways that they couldn't before. So examples we hear are, Hey, I don't really feel comfortable asking this in class. I'm worried that my peers or even my professor might judge me, but I can ask it in guided learning.
And I feel like it's this safe space internally even that we had when we were. Testing the product before we launched it, we had internal colleagues who were using it and they said, you know, I had these things I wanted to learn, but I always felt nervous to ask, but I felt so comfortable and so psychologically safe that I could ask anything.
I was really encouraged and there's that really soft side to the conversation that we wanted to get. Right. How can this be a place where. You can ask anything, you can learn anything. And it's there to really encourage you on that journey in a way where like it still is a student choice. So like the library example I gave students are like, yeah, sometimes I do just need a quick fact.
I forgot something. I wanna check it. Great. Guided learning is an option right in the toolbar so that when you do wanna learn, it's just a tap away. You have an awesome learning experience. But that choice in creating this space where you can learn what you want, you can ask any question, is really like.
The heart of guided learning.
[00:19:33] Alex Sarlin: Yeah, it's a really good metaphor, and I think psychological safety is also part of Google's DNA as a company. That's like a famous aspect of Google culture, so it's an interesting way to think about it, injecting that into the learning product. And one of my favorite instructional design concepts, it's always stayed with me, is this concept of the expert blind spot that, you know, when somebody's an expert in something, they almost forget how to break it down to a novice.
Even realize they're using complex jargon or that they're assuming all of this understanding and they're trying to explain something in a basic way, and they're assuming all this understanding. So when you talk about students being able to say, I would be embarrassed to, you know, interrupt my professor and say, I have no idea what you just said, or like, you just use three acronyms in a row, and I don't know what any of them stand for, but the idea of going to guided learning and being like.
This is what I'm dealing with. How can I unpack this? What does this stand for? How can I put it together? How can I make sense of it myself? It feels like a really valuable use case. I think it's a big gap in a lot of teaching environments.
[00:20:29] Dave Messer: Totally. I think it's that idea of giving people agency over their learning, like they now have the tools to personalize their education, right?
Like you can ask those questions, you can get help that bridges the gap. And I think one thing that's really cool too on the technical side, but you can bring in like your lecture notes or whatever you have written down, and literally make those connections. Like, I wrote this but I don't understand. Or I can say.
Like quiz me on this. I can check my understanding. Like all these features that I actually have in Gemini now that are so helpful for students, you can get that help like so easily, which is incredible.
[00:21:05] Alex Sarlin: Yeah, I think, I mean, making that connection, being able to take your professor slides or your own lecture notes and turn it into something that's interactive that you can use to study and make sense of is a big part.
I think a lot of the AI tools, and I think it's a core, core idea for guided learning. It's really exciting and notebook as well sort of designed for that as well. So you mentioned how when you talk to students, they're so creative and how they use guided learning and how they use LLMs to personalize their own learning experiences.
I'm curious if you have examples of design choices you made for guided learning that came directly from feedback from either students or educators.
[00:21:39] Dave Messer: Yeah, definitely. Even just the existence of guided learning. We were talking with students. This one student, her story, sticks in my mind. She was explaining to me I was using this tool to help me with my math homework.
It just gave me the answers and she said it was great for a little while, but when the test came, she said she bombed it and like she actually stopped using the technology because she realized like, Hey, this has like a long-term effect on my learning. I don't wanna use it for that. And so I think that's what guided learning is about, like how could technology actually 'cause it will answer questions that you ask, but if you wanna understand it and go deeper, like that is right there for you.
So even having something that encourages longer term learning and like helps you really retain information in a way that you understand it, like that is the heart of it. So for her, like. Well this is something that she can use now to like study for, you know, when she has homework, she doesn't understand it, can understand it better, can apply that on the test.
So that's like one example, right? Like I think the heart of it is, hey, when you want to learn, this is a great and easy way to come in and learn. There's a lot of other things that we heard from students who around. You know, they wanted it to be a choice again. So like they explained that, hey, you know, don't make this the only mode I could be in, right?
Like, I don't only wanna be learning, there's sometimes where I need other things. So even that influenced the way that we shaped it, the behavior of it. We got, you know, so many students got to test it and give us feedback on what was good and what we could improve. All of that feedback helped us to understand some of the nuances of like their learning journeys and then what we could do to make it better in those moments.
[00:23:15] Alex Sarlin: Yeah, it reminds me of the metaphor you were using earlier about the librarian, right? It's like sometimes I go into the library because you need to look something up and find something specific out, and what you're really looking for is retrieval of information that you can walk away with and use right then, but in other times, that's not at all what you're looking for.
You're looking for an idea to. Moved into your head to be something you, you, you to be motivated to change your schema, to understand the world in a different way because of something that is in the library. And you don't want the librarian just to hand you a book. You want them to sit down with you and say, Hey, what class are you in?
What are you learning about? Oh, it's that, what's. Discuss how this is relevant to you and your goals and what's interesting to you. And here's what's really cool about it, and let me share my passion for it. It's just like, it's such a different relationship when you're trying to build hooks and retain information and be able to retrieve it later versus just handing it to you.
But it makes sense that students would want both because both are important, they're relevant. One thing that I think is so interesting about guided learning as a product, one of the things that it cues you to do, which is exactly relevant to what you're saying. Quiz is to practice to retrieve. It says, you know, Hey, you want me to make a quiz for you?
You want me to make some flashcards and we can review this together and let's see which, what? What parts of it. You remember it jumps in and tries to suggest some learning modalities at times, and I'm curious about how you think about that from a design perspective. You know, I can imagine some students are.
Yeah, of course I want that. I'm preparing for a math test and I wanna make sure I remember this. And others might be like, oh, I, I was just asking, I don't need flashcards right now. But you have to sort of figure that out in real time with the intelligence of the model. How do you think about that moving into a slightly more like real tutoring mode versus conversation mode?
[00:24:56] Dave Messer: Yeah, I think it's similar to how I would think as a teacher, right? So give students choices, give them some agency in their learning, and when there's different possibilities, they'll pick the one that's the right fit for them at that moment. So like no one is required to use guided learning, right? Like you can choose to use it, you can use it when it's helpful for you to learn.
And so I think it's more of a tool that can really like offer different. Things to learn, different ways to learn, different ways to check your learning. And in the moment when you get those invitations, you can take them up and say, yeah, let's review. Like, I actually wanna see what I learned and check it other times that might not actually fit.
So like, you know, my example of how do animals breathe, right? I don't really care about being quizzed on it 'cause I was just curious and I kind of wanna understand it. Yeah. But if this was something that was, I knew I would be assessed on, or I cared about making sure that I got it right and, and understood it more deeply.
Yeah, that's amazing. To actually have it say, Hey can create this for you. You can check your understanding with the quizzes too. I think one of the really powerful things there is it can be grounded on your conversation and then again, anything you bring in. So if you're, I dunno, as a teacher I didn't always have study guides or practice quizzes.
I was trying to just make like the one really good quiz for the class, but that students can say, Hey, yeah, let me actually go and like. Take a quiz, see what I've got and what things I need to work on. It can give you this experience, give you that personalized feedback. So you know, like, okay, I need to study this a bit more and actually jump back into your learning loop and like understand those things go deeper.
But I think having those tools available, Gemini and guided learning, suggest them when it's helpful and then letting people choose. I think that's really the heart of kind of how we approach it today.
[00:26:36] Alex Sarlin: Totally agree. I mean, I'm a destructional design background and I know that practice quizzing is pretty much the most valuable thing you can do if you wanna retain knowledge.
It's like absolutely core and doing it in a spaced way and interleaving, and there's all these techniques that are proven, or at least they have good evidence for that. They help people retain knowledge, chunking, there's just all this stuff about that. We talked about learn a little bit, but these pedagogical principles, can you break it down a little bit?
I think learn LM has these five. Criteria on which it sort of measures itself, or like metacognition, I don't remember exactly what they are, but maybe you can sort of unpack a little bit about what it actually means when it brings the learning science into the conversation, what it's trying to optimize for.
[00:27:13] Dave Messer: Yeah. Okay. Those are great. So I actually have these handy 'cause they're super important. Yeah. Inspiring active learning is our first. We really wanna take things, again, not from like passive consumption of information, but actually getting you to think, getting you to learn and engage actively. That's where the magic happens.
It's where you make those connections. It's where you like, you know, really learn managing cognitive load. So that's another thing you'll see in guided learning where like. It's not the wall of text you, it really chunks it down into like smaller bite-sized pieces, if you will. And then even bringing in the multimedia helps there too, where it's a lot to read a really long description of something.
It's easier when that's paired with a helpful visual that you can engage with. We wanna deepen metacognition. We think that like. Learning isn't just about the information, but it's actually like learning how you learn and thinking about that, reflecting on it. That's a key part as well. A lot of what guided learning will do too is ask you questions and like ask you about your learning, ask you what you think.
Those are key parts there. Stimulating curiosity, like actually getting you engaged in the topic. I think that's so important 'cause learning isn't just about, again, the information, but it's about bringing it to life. It's about asking more questions and getting curious that you wanna learn more.
Adapting to the learner is a big one, and this is like personalization, you know, broadly or even like in a conversation adapting. To their needs, what's helpful to them? I think that's a really big piece. And then actually a new one that we have added recently is preparing learners for the future. So how do we think about those, those core skills and competencies that will help people? And actually, you know, in the work that we do, think about ways that we can prepare them for their future, whether that's for work, for life, for all those different things.
So those are some of the different principles. And then. Again, you'll see those in our work we do in our model with Learn LM and Gemini. You'll see that in our products across search classroom, YouTube, Gemini. Really, these are our like ways of us looking at the experiences we're creating, making sure that they're meaningful and informed.
By learning science and ultimately helpful based on what we know is helpful for learning.
[00:29:21] Alex Sarlin: It's a powerful model, and I think it touches on so many of the really important aspects of what good learning looks like. It's active, it's metacognitive, it's personalized, or it's adaptive. For the learner, it makes a lot of sense.
It's based on real learning science. I wanted to double click on the adaptive personalized piece for a moment, just because this is such a rich area for ed tech right now, and I think a lot of our listeners who are ed tech founders, ed tech investors, people who just follow the space really closely have been wrestling with this concept of individualized, personalized, you know, adaptive learning.
It's been a dream of the education technology, and education. Proper community for a long time to be able to deliver learning that is not one size fits all, that is adaptive to how students think or what their preferences are. All sorts of things, and I feel like we're closer than we've ever been. I truly feel that.
One thing that I think is incredibly interesting about Google, but it can be a little bit of a double-edged sword, is that Google has so many surfaces, right? I mean, there's so. Ways in which students and educators interact with Google all the time. They're using mail, they're using maps, they're making slides, they're using docs.
You can just keep going. They're in classroom, they're watching YouTube, and there's a world in which different aspects of that data could be used to personalize learning as well as previous chats in guided learning. You know, if you come in one week and say, I have a math test next week, and then you come in the day before the test, you know, it should remember that you have a math test.
The next day. There are all of these ways in which that data has the potential to truly be adaptive and personalized learning. If you know that that student, all they do all day is watch soccer videos, or all they do all day is watch, you know, crafting videos, well that's a really good signal about the kind of thing they like doing and something you might wanna use in your tutoring session.
But it's also obviously. A data issue and it, you have to be really careful about where and when you're collecting data or moving it around. I'm curious how you think about this, and obviously this is not meant to be a, you know, spell out any kind of roadmap or data policy, but I'm just curious how you think about the data that's being gathered on any individual student or that has the potential to be gathered and how that could be used for personalization in a safe and secure way.
[00:31:23] Dave Messer: Yeah, that's a great question. I'll start by just talking about Google and our privacy principles. Privacy is so core to what we do. Making sure that you know, for every product to use, the privacy policies are transparent. You actually have control over what data is shared, how it's used, that is like core, so that user trust, making sure that people.
In control of that. They know it and they're choosing what to share and when is so important. We have entire teams who we work with closely to ensure that we're doing the right thing for the user and making sure that we follow those principles. So that's first and foremost. When I was working in schools, actually did a lot of work with personalization as well, and like at the time what that meant was.
Use a lot of ed tech tools, get the data, we'd analyze it offline. We would develop plans for what to do. We would make these instructional plans and like these, these cycle times were really long, right? That's a lot of work. We have to sit down as a team and crunch the data, make decisions, make some like, you know, four to six week plans and like that was, have broader personalization and differentiation and then really leveraging like a lot of platforms that were more.
Personalized to students too, both in like the instruction and the feedback where you could get feedback with those things. I think with ai, like that loop could be much. Tighter, right? Where like actually, you know, understanding what you need, understanding and finding or creating a way to learn it.
Engaging with that material, like you can imagine very like that ability to personalize, learning to, to understand what someone needs and then have great, helpful learning experiences tied with it. I think that, you know, technology. Definitely help with that. I think one other thing that I'm passionate about is making sure that that heart of the classroom, the teacher, the connections, the community is amplified there.
Like how do you not lose that? Because I think on one extreme you have a bunch of students just staring at a screen all day, and that's not the future we want. I think the one we want is the one that we wanna help create, like the kinds of tools that. Bring us toward what we want and like how do we actually create flourishing classrooms?
Ones where like those, those connections, maybe it's actually helping a teacher 'cause they get the insights about where a student is stuck and then like that helps inform decisions they make about what to do. But those cycle times, it doesn't necessarily mean it's all a student with a computer, but how could those cycles and capabilities be brought into the way that the classroom works, the way that students work, so that it's helpful.
I think like data is helpful for sure. Where. Again, if users say, Hey, this is something I want to share, and there's benefit to me, that's amazing. I know in Gemini today you can integrate your Gmail and your calendar. I use it all the time. Like we'll have like the sports schedule for a whole season.
I could just take a picture of it and say, add this to my Google calendar. It's amazing and like. Me connecting that data. Awesome. Like I'm in control. I choose it, but the value is incredible and what it can do is great. To your point, when you think about different ways that people use Google products, maybe there is some interesting things that are helpful there too, that if you could connect the dots, maybe you have this platform that helps personalize across different surfaces.
I think ultimately what we wanna do is focus on the user. What do they need and what would be helpful. And there's other ways of doing this too, right? Like a very simple thing we could do is just ask them, Hey, if guided learning says, hey, like are you studying for a class? Do you have like notes? That might be enough too.
So I, you know, I think there's different ways of solving the problem to personalize your learning. I think we all kind of go to this place of, we have to know everything and all the data, but you know, there's also other ways of doing it that. Or simpler and still helpful. So I think there's a gamut of options.
I think that we wanna make sure that users are in control, they choose it. But if there's opportunity to be more helpful, that's amazing. And we would definitely want to pursue things that help people learn and help people to achieve even better.
[00:35:18] Alex Sarlin: That's a great answer and I appreciate it and I think it's totally makes sense to focus first and foremost on data privacy, on opt-ins, on clear data, you know, silos or you know, being careful about data, especially student data when you're talking about things like Google classroom data that undergirds absolutely everything.
But I love your example about integrating, you know, the sports. You take a picture of the sports, uh, calendar and it integrates it into your calendar because you can easily imagine how valuable that would be for a, you know, a student who has five different classes they're taking in a high school or a college, and they can take a picture of the whole, all the assignments in the calendar and suddenly have an academic calendar.
And if they're guided learning tutor knows. That then it can remind them when to study and give them a whole study schedule, which could also go in the calendar. And I mean, that's just one of dozens of different Google tools that could integrate, but it's so powerful already. I think the learning preferences piece, the like, oh, they watch videos on X, Y, Z topic on, on race car driving, so we're gonna make their.
You'd make all their tutoring about race car driving. That's a go-to example that like I think everybody goes to that may not be the most relevant, even though it feels sort of appealing at some visceral level. But things like integrating with your calendar that makes, just makes sense, you know?
[00:36:29] Dave Messer: Yeah. I mean, just on the example of like race cars, right?
I think the idea that your entire education is based on a niche interest you have is probably not a good thing. But I think where it's helpful if you think about your time as a teacher, right? If you knew a student was passionate about something, they've got this whole mental model of that space and they probably really understand the domain.
And so why I think it's actually really valuable is like you draw on that to make the connection. And so anything that connects with them and you can bridge that gap, like that's the magic. So finding those opportunities, finding ways to make connections to. What they're interested in, their community, things you've learned before.
I think that's really the art of what it's about. And so like even in Gemini, having remembered past conversations to draw and I think that's like even a, a really powerful thing.
[00:37:16] Alex Sarlin: It really is. And even that is a relatively new feature for all alums that's not like, hasn't been baked in since the beginning.
We're getting on time. This is such an interesting conversation. I feel like we could talk forever, but one of the things I wanted to ask you about is, you know. We are in a really interesting moment. I think, you know, we follow this pretty closely, the sort of sentiment around ai and I think people are both very bullish and very excited about the potential for ai.
And there's still trepidation and there's still fear. Uh, you mentioned the sort of fear of cheating a while back, which is definitely real. One of the other. Sort of aspects of AI that we've been starting to see is that a poll just came out recently in which a teacher's support of using AI in the classroom for various things has went down a little bit year over year, and parents were expressing a little bit of concern about their students' data, basically, or AI tools, getting access to their information about their students and.
We're obviously in this tricky moment with technology in general. People are both trusted and use it all the time and feel like they should be really careful about certain things happening because of breaches in the past from certain places. I'm curious, when you talk to educators and you say, we're doing guided learning, we're doing this amazing project that it allows, you know, students to, to learn anything they'd like, it's pedagogically infused.
What are some of the concerns that you hear and how do you assuage them?
[00:38:32] Dave Messer: Yeah, I really appreciate it when we step back and kind of look at this moment. The technology is opening new questions 'cause like its capabilities are new. I don't think we've ever grappled with what this means. Google's approach, you know, is really to be bold and responsible in this space.
We want to again focus on users, what's helpful and how we can continue to push forward. But taking those concerns and understand what we could do to help with those. Guided learning is even one example of that, right? Like there's concern about cheating. This is a way we're actually showing you can use AI to learn.
I think when we talk with educators, what's amazing is actually really excited about guided learning and they have more requests like, I needed to do this and this and this, and then, you know, so the technology is still early, I would say, and there's a lot more work to do, but. Understand what people need and those both the opportunities and the concerns.
I think that the magic is one we can solve for both. Find the way to move forward that balances them. You know, teachers that are seeing like, Hey, this is what I need with the tool. Like that's awesome feedback. We've gotten that feedback from folks. And those are the kinds of things that help inform the work we do in, in the ways that we can continue to make it better and more useful.
I think the conversation that we're having about the value, how it can be helpful, how it can be hurtful is really good and healthy, and we should keep having that conversation so that we can find the right solution together and continue to make things better. I believe in the potential of the technology to be used for good, to be helpful for learning and teaching.
I totally acknowledge the ways that it can be used in ways that can diminish that. Like if I'm just outsourcing my learning to ai, it's not gonna help me. But thinking of what are ways that BA basically help with that, right? So whether that's guided learning, whether that's, even Google has done a ton of work just around AI literacy.
I think a lot of times our own relationship with this technology is the biggest thing. How you use it, how you know how it works. Like those things, I think can influence decisions for learners. I think the really cool thing is you actually are like confronted with this choice about how do I want to learn?
And you can either go the route of just gimme the answer or how I wanna learn. And what we're super excited about is like how do we continue to make it really easy for people that want to come to learn? How do we make that an awesome experience and continue to work with educators and students in ways to build their literacy so that they understand both how to use these tools in different ways that they could use them.
There are cases like students are doing amazing things. They want to learn. They're like. One student shared this example where when they don't understand something, they ask for it to be explained as if it was gossip, and they're like, this is so great. I love it. And I like get really engaged with this topic.
I like, it's so cool and students are so creative. So I think there's both sides to this. I think again. The more that we can help create tools that are helpful and then can help in the community, both listen and then help people understand how the tools can work so that they can use them in their lives.
That's really where the real like value in magic is.
[00:41:32] Alex Sarlin: You mentioned that students want to have multiple modes. They wanna have guided learning mode, but they also wanna be turned it off if they want and just ask questions. Do you have educators saying, no, they should only be in guided learning mode. Is that a request you've gotten from educators?
[00:41:45] Dave Messer: I'll say like some, but I think there's so many tools that are just like amazing. Like an example last week we did Gemini is we were like at Notre Dame and like talking with students and faculty and one of the things that the faculty were sharing is actually using AI as a way to help people learn. There's a student who is like. He was really deep into history and like this really niche topic and he told me, you know, no AI gets this right. And we used deep research and we went so deep on this topic and got an amazing overview. He literally like shouted for joy because he's like, this is great. But I think it's again about like what can the tools do?
How can we use them in different ways? Yeah. And then ultimately I do believe like our relationship with what, what we want, how we're thinking about our own learning and the choices we make when we use the tools like that is the key part. So I think like, yeah, in short, like yes, people have asked to like, you know,
[00:42:54] Alex Sarlin: but I think it's like a lot more nuanced.
No, it definitely is. It's a great answer. And I think Agao example is such a good one because it's like if you give students autonomy, you give them agency. You don't assume that they're gonna use it for nefarious purposes. They do all sorts of really interesting and amazing things, and I think we should remember that and not always get into this sort of arms race mentality.
I always complain about that feeling and when it's like teachers are trying to stop students from using AI and students are trying to stop teachers from using AI. It's just, it's ridiculous. This has been such a fascinating conversation. Dave Messer is a product manager on Google's learning and education team.
He leads product for emerging learning initiatives, including Learning Labs and Gemini has been working on guided learning. Thank you so much for being here with us on EdTech Insiders. Awesome. Great chatting with you, Alex.
Thanks for listening to this episode of EdTech. If you like the podcast, remember to rate it and share it with others in the EdTech community.
For those who want even more, EdTech Insider, subscribe to the Free EdTech Insiders Newsletter on Substack.
