The State of AI & Education

Justine Moore: Will software in some way replace teachers? Parents want better outcomes. What is actually improving? Retention of information, learning, memory. They're allowing a teacher to be a lot more productive and be 10 times better at their job. Will we see more parents like, "I want more of a truly AI-directed education for my kid?" It's teachers who are willing to pay and use this in their every single day workflow. We need the textbook companies to grow some innovation arms fast. AI is already reshaping the way students study, teachers teach, and creators deliver knowledge. But the bigger question is, are we on the brink of a complete rewrite of how learning works? On today's episode, we're joined by three A16Z consumer partners, Olivia Moore, Zach Cohen, and Justine Moore, who together bring perspectives on the cultural, technical, and market shifts transforming education. From TikToks featuring AI-generated celebrity tutors to private schools running full-stack AI curricula to the surprising way teachers, not students, are driving adoption. This conversation discusses what's real, what's hype, and what the future of learning might look like when it's built natively for AI. Let's get into it. As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16z.com forward slash disclosures. I'm Justine, and this is Olivia, and welcome back to This Week in Consumer AI. This week, we're going to be talking about a very fun trend we've seen recently blowing up all over our social media feeds related to AI and education. And for this topic, we wanted to bring on our subject matter expert here at A16Z on education, which is our colleague, Zach. Zach, you want to give a little introduction and talk about your experience in ed tech?

Zach Cohen: Yeah, I would love to. I think it's fun to feel invited to a podcast from colleagues. It's an interesting feeling for sure. Yeah, so quickly, just my background. I've been operating and investing in education for quite some time. So I've been here for a little over two and a half years, thinking about what the next generation of education is going to look like with the advent of AI. And then before this was at General Atlantic with one of my coverage areas on the consumer internet group, being education. We were investors in Quizlet, Duolingo. We did the chess.com deal while I was there. So the edutainment kind of consumerization of education. And then before that, I actually built and sold an education technology company that was focused on delivering high school computer science education in mostly the Northeast area and we sold that to actually an education-focused roll-up. You can count how many times I've said education, where I worked for a couple of years as well, rolling up different assets in mostly adult education and corporate training. And I know we have looked at a lot in ed tech over the past two years, especially. I think everyone has this intuitive sense that people are using AI for homework. But even I was surprised, like seeing some of the data for probably the full first year of ChatGPT, like the .edu users were even larger than everyone else who was a non-student user, which is pretty crazy. And it does feel like the backlash was also as severe and as immediate, like all of the anti-AI checkers. Are schools still anti-AI? Like, how has that evolved?

Justine Moore: Yeah, it's a good question. We had like AI wave, then AI detectors, and the rock'em, sock'em wars of education. And you had the LA and New York City public schools banning AI right away. I think we're super far away from that moment, which I think is really, really strong. Now, I think we're in different layers of far away depending on where you are in education. So if you look at K-12, there's still some skeptics. There's still, I would say, like predominantly, like I think 80% of districts now have a generative AI team that's going after and looking at procuring kind of new technologies. So not only is there like earmarked budget, there are people who are thinking and proactively looking for it. But there's still friction around what you want to use it for and how you want to use it. And maybe there should be a teacher in the loop that has more understanding of how students are using the AI. And maybe that's a good or bad thing. But I think it's a good thing, kind of net-net, that there will be more AI in the classroom. And then higher ed, like you have Claude releasing Claude for education. You have OpenAI releasing kind of education platform. And they're partnering and piloting with a bunch of universities. And this is like a pretty horizontal platform that will be like fundamental to teachers and students' experiences. So I think higher ed is leading the pack. I think they're realizing that this is going to be a tool that people are going to need to know how to use in their jobs, in their everyday lives. So I've been pleasantly surprised. I know, you know, 18 months does feel like a long time, but in education, it's not. We've taken five, six years to move to cloud and like most people aren't even there yet in the education world. So I think it's super impressive. And there's some schools that are like mandatory AI usage, like in the curriculum. I think Ohio State is one of them. So I think it will continue to trickle down. I think people are always a bit more careful and wary about technology with the younger ages. But I think we've passed the kind of hysteria moment, are now in kind of the pragmatic moment, which is exciting, especially for founders building a space.

Olivia Moore: Totally. For sure. Yeah. And I think you sort of alluded to this when talking about your experience in ed tech. But one of the things that's so both interesting and scary about the education market as more outsiders is like it's not just one industry, right? There's public schools. There's private schools. There's charter schools. There's homeschooling. There's supplementary products that parents buy for their kids. Then there's the whole higher ed, public and private industry. And then there's like adult education, kind of future learning, reskilling, all of those sorts of things, which I would imagine is probably going to be huge in the age of AI, automating many jobs and or enhancing many jobs. I'm curious, have you seen the most adoption among a segment of that market, like parents just buying standalone products for their kids versus school districts?

Zach Cohen: I'm still really surprised that this is my answer, which is teachers. It's not actually students. It's teachers who are willing to pay and use this in their every single day workflow. I've seen a lot of adoption at students who are trying to use this for homework helpers, but they're turning off or using it as a way to get their problem set done and then go back to kind of hanging out. And adult learners, like the retention maybe has been a bit better or like the completion rates or learning efficacy on Duolingo has improved with all the AI features. But I have not seen a native AI player in the adult education space. And maybe that's just because distribution is really hard and layering on AI to existing strong pedagogy is maybe the right way to do it. Or it seems like that is the case now. But with teachers, like the rapid amount of adoption is so high. And that's a lot of it is because like 90 percent of the job that they hate is the administrative part, which is grading, feedback, going home, building new assignments, new curriculum. A lot of them have just been borrowing curriculum year over year. They've been trying to iterate on one unit at a time and now they can generate a curriculum per student. And I think that extremely exciting. I think we hopefully will take it farther and it won't be generating assets that we already seen in education like worksheets or multiple choice questions. And like the units will start to become more of an AI experience. But for now, it's been teachers the wave. I think Magic School has over like 5 million users. I think 50% of US teachers has used their tool. Most of the most mature AI companies from a revenue standpoint have been selling to teachers bottoms up, which is tremendous. Teachers have a very small wallet to spend on tools, and this is just an outsized return for them compared to the 15 or 20 bucks a month that they're spending on it.

Olivia Moore: This brings up another point when you think about what's driving student outcomes, which is like, there's a couple ways to look at what is working in AI and ed tech. And one would be like, what is getting the most usage, which to your point is probably still ChatGPT for the end students versus like maybe more focused products. And then there's what is actually improving retention of information, learning, memory, things like that. Have you seen anything that you've been really impressed by in terms of what's working in AI and ed tech?

Zach Cohen: Yeah, I think the core of the question is the question, which is like, what is working, which is like, what does working mean? Yeah. There's the investor side of what it's working, which is like, what do you want to see when you're investing in a company, right? Which is different than like the learning outcome side. I think from the investing side, I care a lot about looking at kind of retention and engagement. Retention, obviously, because you have summer months, but you also have students who come on with a test that they have in two days. And is the experience good enough that they capture them to work proactively on the app the next time? Right. So we look at metrics such as cohort engagement to see how students are performing, things like that, or monthly retention, less so like the weekly or yearly retention, which I think could be kind of a red herring in this market. So I love looking at like how many days in a week is a user using this on a cohort basis as the company grows, is this starting to become part of their learning? And I think that that is also a good proxy for how good the product is, not just how engaging it is. Because if a student is doing this to just get their homework done, I don't think you're going to see that level of engagement. You're going to see it taper off around certain times, certain months, or be a little spikier. But people who are seeing flattening engagement of number of days per week, hopefully, you know, about four or five given the school week, I think it's exciting. On the education side of what's working. This is a really, really hard question. I think it's probably, like a lot of different markets are kind of suffering from this, which is like, what is the benchmark and eval function in education, right? So luckily education, like it's not like a corporate job where it's like, oh, you have an ambiguous performance review and maybe you're doing 10% better. There is like tests at the end of the year that measure students' performance. The problem with that is you need multiple years of student testing because there's a bunch of variables there. So one is like, how do you figure out like what's a conditional, what's independent, what's dependent, all of that stuff in terms of like what's working in the school. But on top of that, AI is still at the periphery of education, right? So since it's at the periphery of education, we don't actually know what the outside's impact is. It's helping teachers create assignments better. It's helping students get homework done, but it's not like core to the schooling system. So I think it's really hard to know learning outcomes. I think it has to be a mix between like the metrics of the product and how good an understanding the founder is of like how to build kind of pedagogy and learning methods within the product. But I think we're still like a couple years away from actually understanding like, you know, what the learning goals are. There's some papers that have come out from various schools and universities have shown like test improvements with an AI instructed course or things like that, which is exciting. But again, these are research papers and case studies. These aren't kind of like full state or full nationwide studies.

Justine Moore: Okay. So to get back to the point about making AI core to the educational experience that kind of touches on one of the sort of viral social media accounts we've been seeing. And I'm sure anyone in the education space, definitely you, Zach, has been seeing. I think it's literally called The Future of Education. It's an Instagram account. It's run by this woman. I think her name's Mackenzie Price. And she runs a school called Alpha School, which I believe is a charter school in a few states now and has had these viral Instagram videos blow up by showing how her school, they have two hours of classwork, I believe it is, a day, and it's delivered via AI tutors, like these computer programs that are teaching kids at their level. And then they spend the rest of the day on sort of self-directed projects or things like that. Would love to know your take as the education expert, Zach, what you've heard about Alpha School or will it work to deploy AI like that?

Zach Cohen: Yeah. Alpha School is like everything I can ask for about what's happening in education. We talk about it, we opine on it, and then we have to just sit here and wait for superintendents and district heads to buy single point solutions in AI every turn of the budget. And then over three to four years, we can have a narrative. I think what Alpha School brings us is what like any labs team would be a bring at a company, right? So it's like, hey, you know, the Snap Labs team is going to try out a bunch of new features and they could deprecate whatever features because they're moving fast and not deploying a bunch of resources and they have outsized budget to go do this, then let's see what works and what works we're going to take and what doesn't work. We're going to deprecate and we're going to continue to iterate. I think Alpha School is doing that for private schools now and hopefully for public schools later, which is a tuition is, you know, $40,000 or super well-funded private school. The parent and student base is self-selecting into a group who wants technology in the classroom. These might seem like just small facts, but like all of these in amalgamation change, like the outcome of what something like a school can do and removes a bunch of friction to allow you to move fast. So I think what Alpha School is showing us is if you try to turn full tilt on AI and education and you bring it into the classroom and you're fine overspending on software, you're fine spending a bunch of money testing a bunch of different softwares, you can figure out what's working. I think my takeaway here is this is a great signal for where technology is today, that if you have the privilege to go test that and figure it out, it will have outsized impact. I think they rank like 99th percentile in a few different assessments. Their students are now like top one, top 2% of the nation. Like it is incredible outcomes. But again, this is like the lack of commercialization of a lot of this software. Part of that is what does the integration look like? Part of that is just the cost of software. Part of that is like the software literacy at the school level. So I think Alpha School is net-net extremely positive signals. I think it's an incredible way to learn. It's leaning into a lot of the old ways that we've talked about learning, which is like self-discovery, self-learning, very limited instruction. And it's working. So I think public schools are looking at it. Maybe superintendents are saying, well, I don't have hundreds of thousands of dollars of a software budget. But I think that's okay, right? Like we see software gets cheaper. Let, you know, the Alpha School as a world test out what's really working. And those companies will really benefit from early partnerships from Alpha School. And to me, it's like paving a way for a lot of companies to see what they're doing. Obviously, from a very privileged standpoint where Alpha School is.

Justine Moore: Interesting. Yeah. I think the controversial question there, like when you talk about superintendents don't have hundreds of thousands of dollars to spend on software every year, they do spend that on teachers. And so the controversial thing is, will software in some way replace teachers? Will AI replace teachers? Or will we still need a human instructor to be delivering the curriculum or giving lessons to the majority of students?

Zach Cohen: Yeah. I think no or very long horizon away. There still a shortage. There still teachers who are swamped with work. What we talking about in the AI world for what consumers can do with AI and what even enterprises are doing with AI like we are multiple years behind there. Like the biggest use case of AI right now is generating an answer sheet or a worksheet assignment. It's not even like a unit in the classroom. And what I mean by that is like it's actually a really big nuanced difference. And one of the reasons why I've had a little bit of a hard time investing in these early education companies is they're allowing a teacher to be a lot more productive and be 10 times better at their job, feel like burnout is a lot lower. Incredible, right? Really, really amazing. But it doesn't mean that the students are interacting with AI in a learning environment. They're actually interacting with the same asset that they would, and they're having the exact same experience. Maybe they're having updated worksheets. Maybe there's some like better memes or better references in those worksheets because the teachers can update it weekly. But that's very different than being in an AI enabled environment where you can take a history lesson straight from Abe Lincoln's avatar and speak to them and have a detailed conversation or create your own world around what you think. You know, you write a creative writing prompt and go create your own world around it or create a game around it. Like I think those experiences are far away from coming. So I think my point here is not to not answer the question, but to say we're so far away from even like AI teaching units. I think we're very far away from AI teachers. Now, I think what will happen is like the amount of active teaching will go away because I think you can like have really good active teaching with AI, but I don't think it's going to fully replace a human teacher.

Justine Moore: Yeah, super interesting. I do feel like on the opposite end of the spectrum from Alpha School, which is like very private, very high cost, is this explosion of like free, publicly available, what I would call ed tech content, but it looks very different from the prior best in class, which was like the Khan Academy videos that I learned math and other things on and that I love. That open source animation package is called Manum. It's still around today. Yeah. And now we have like people putting their textbook into Notebook LM and making a podcast. You found a bunch of TikTok videos that generate like Sydney Sweeney explaining this. Yeah. There's this Instagram account called Onlock Learning. Okay. Yeah. And every video they post gets millions of views. And it's like explanations delivered by like deep fake AI celebrities of like math and physics content. I think particularly for the AP or IB exams. Yeah. And it also has like diagrams and graphics that sort of walk you through the explanation. Yeah. And I guess it just makes me think to what extent are we going to see standalone AI ed tech products versus just new formats and ways to learn that are delivered across all of the platforms that we already have?

Zach Cohen: Yeah. I think it's a great question. And I've watched way more educational videos than I have in the last like four days than I've had in the last 10 years, which again is like a really good signal that like they're engaging. They feel like brain rock content, but they're actually the complete opposite. Yes. So they like look, feel, sound are cut the same way. But they're teaching a really, really, really, really detailed topic and pretty technical topic. And like they're getting better. Like I remember a year ago, it was like Taylor Swift teaching the Taylor series. It's a tool we use in math to turn scary functions into friendly polynomials. and now you have like Drake in conversation with Sidney Sweeney so if we square root it we get the vector's length and that guys and gals is why this expression gives you the length of a 3D vector yeah talking about three dimensional shapes it is like getting better both on like the animations of the graphic side and then also the deepfakes of the celebrity side like it's all getting a lot more engaging what I think is really interesting here and the thing to focus on was I don't know if you remember as a kid you were like oh like Zach's a visual learner or like Justine likes audio or Olivia likes to read and like marinate with her ideas. And then you kind of get boxed into like a type of learner. And that was considered like pretty progressive education. It's like, oh, we know what type of learner you are and we're going to like cater to it. The problem is I'm a visual learner for one thing, but I might want to listen to a podcast for another and read for another or watch a brain rot radio for another. Like it just could be so many different things or to do a hundred problem sets. Right. So I think what's happening now is like we're going to have like a factioning of a bunch of different types of learners where depending on the topic and your understanding of the topic, you could pick whatever modality you want. And I think that's really, really exciting. And then also depending on how serious you want to get in the topic. Like if you want to just be conversational around it and it's not for a test, then like maybe brain rot content is perfect. But if you are taking like an exam that matters for your final grade, like maybe you want to generate a bunch of problem sets and then listen to audio that explains your answers, right? So I think for me, it's more, you know, I don't know if any of these turn into individual companies, but I think what's really exciting here is like the mode of learning is going to be like the title of what mode of learning you are and what mode of learning, like you're going to be like burdened. You know, it's going to be removed from that burden. And I think that to me, that is like the most exciting aspect for education right now. And I think like we're seeing it, right? The engagement of these videos are insane. I mean, this is not a brain rot video. So, you know, this is not Italian brain rot. Like this is very, very different. Yeah. Well, even the VO3 vlogs where it's like a day in the life of some famous historical figure. Yeah, the history stuff is incredible. It's fascinating. It's like truly entertaining to watch. And I find myself like actually learning things. I think the hard part of that is there's a few different companies that are doing really interesting things at the forefront of AI and education. Yeah. One of which feels like what the VO3 thing is doing, which like you can chat and have a conversation with Napoleon about history, right? Things like that. But in order to get school adoption, they have to wedge in with these kind of like more rudimentary tools, which is like the worksheet generators and like student helpers and feedback. And then once you do that, it's not so easy to get people to use the other mode because it's a new form of education versus, hey, do the same thing, just do it 10 times better. So it's interesting to see like the stuff that engages us the most and gets me the most excited. even if it's being adopted and paid for by the school. If you look at the usage of those tools compared to the usage of kind of the more rudimentary tools, it's extremely low. So I think there's probably going to have to be like a large like PD movement here on like how to bring AI not just into your workflow, but into your classroom. And I think that's a really big difference. Yeah. It's fascinating to think too about if you were to design the school system from the ground up starting in the AI world, like how it would be different. One of the things that has really interested me about the brain rot videos is like, for the first time, it feels like it's separating like what the content is and who is delivering the content. And then it's optimizing both of those. So if you look at the comments on these videos, I'm sure we'll show some of them. It's often people who are like, I was like a physics graduate student and I got like a 4.0 at this prestigious program. And I'm like telling you guys, this is like an exceptional explanation and I've never understood this content better. So like you're optimizing for like, hey, this is how you deliver the content in a good way. Explain it in a good way. Make it understandable. And then like, here's the like brain rot celebrity. Like this is a familiar voice. Subway surfers video on top. Exactly. This is an interesting graphic or image to like bring those things together. Yeah. Which we pretty much I think only seen with the self AI products today because the traditional schools you can really separate the content from the teachers is my understanding Yeah it interesting I think the answer is in my head is probably like the textbook companies control this dynamic here a lot. They are still the gatekeepers of a lot of content and what gets in and out of the classroom. Now, I think it depends on whether they run with this or not. I think sometimes they view this as cannibalizing their existing business. Sometimes they're like, oh, this is extending our existing content and we have the single source of truth and AI is really good at augmenting but not net new creation. So I think what's going to happen is really dependent on like where the textbook companies and the publishers spend the next kind of 18 months thinking about net new products. Are they going to partner with AI companies and the AI companies will get distribution, the textbook companies will get really strong product and technology extension of their existing content that feels like it's slowly losing its value day by day. So it's a weird kind of power dynamic right now, but I think that that's where we are. Yeah. We need the textbook companies to grow some innovation arms fast. Or outsource them. Yeah. Like, Zach's on one side of the spectrum, which is this is how the educational system actually works. And it's, like, very complicated. And in some ways, it's like the healthcare system or the tax system where there's a bunch of annoying, outdated stuff, but you can't really change it. And then I'm on the other end of the spectrum where I'm like, blow it all up and have everyone be taught by an AI tutor, like, all of the time. Like if LLMs have ingested the whole internet, they've probably learned a lot of stuff. Yeah. And they might even know what to teach us that teachers don't. Yeah. And I would assume that sort of the Zach perspective is more where we are today. But I am wondering, like, are we going to see a lot of people are basically reckoning what to do with their careers, what to do with their lives, what to do with their education in the age of AI? Yeah. Will we see more parents opting for something more on the second path that's like, I want more of a truly AI-directed education for my kid? Yeah. Parents want better outcomes. Like, that's what the reward model for them is. Are we going to invest in technology? Are we going to invest in this tutor? Are we going to invest in this course that allows my child to be better off from an education standpoint? I think like one of the most interesting pitches I've seen in a company was super early was an AI reading company, which basically like sent kids books. And then they had AI teach the kids how to read. And it was mostly three to four year olds. The promise was to get them to read at like a third grade level within three months. And they had a bunch of parents within the first two months sign up for this. Now, if they can prove this outcome, I think they'll get a lot more parents. If they can't, I think parents are like, well, AI and technology maybe isn't the right way to do it. I'm going to go pay a tutor. I think the other side of this is also if you have the disposable income to spend $100, $200 an hour on a private tutor, the comp to using AI is a bit harder. And if the AI software is priced against that, then you have a little bit harder and more friction on the adoption curve of AI. But if it's, hey, my kid's going to watch Netflix for four hours or I can put them in front of an LLM that can teach them things and keep them engaged, then that's a very different conversation. So I think it's very kind of socioeconomically dependent. I think it's very geo-dependent. I think it's very like resource constraint dependent. So I think there's a future probably for a few students to feel that way, especially because like LLMs can be controlled, right? So like parents could say, hey, I want it to explain this way. You know, if it gets easier to control them, even for non-technical users, which it continues to be. Like, hey, I don't want you to discuss this topic or every question I want you to like show the mathematical equation and write the proof and explain it. You know, I think all of that control is really, really exciting for some parents. So I think we just have to figure out how to harness it and how to benchmark it. And I think it could be a potential future. Looking a year from now, AI ed tech, like, first of all, what do you think we're going to see? Is it a lot of progress? Is it a little bit of progress? And then what would you like to see, ideally?

Zach Cohen: Yeah, so much. In a year from now, I think we'll see a ton of progress on the higher education side. I think this is the year for higher education, this kind of next coming school year. How do higher ed use it? How do the large model companies work with education companies? The question of, are the large model companies going to be good enough? Will there need to be smaller application companies built on top of the large model companies? I think that's one thing we're going to learn a lot of. I think we are going to, in the next 12 months, move from the outside of education, the use of AI just to make things easier to like bring it into the classroom, whether it's like adding it to a discussion. I think a lot of the unlocks in voice in real time is really exciting. So I think the next 12 months will show us a lot. I don't know if it will necessarily like birth a massive company in education. I still think it will move. But I think education will look fairly similar, though. But I think we'll learn a lot as the market kind of matures. Yeah. What I'd love to see from a consumer perspective is the first like fully AI teacher influencer, not one of the deepfake celebrities. I know they're super entertaining, but it would be fascinating to see like what does it look like if you designed the most engaging but also truly informative teacher from the ground up. Totally. And you alluded to this as well, Zach. Like, there's probably not one teacher who's that for everyone. Yep. I might learn best from an animated dog, whereas you might learn best from an actual photorealistic teacher. And how we can create those sort of, like, adaptive AI-driven characters in real time who can deliver really personalized learning on each student's timeline. Because we sort of touched on this a bit, but I think one of the exciting things that schools like Alpha School are enabling is you can spend as much or little time on a subject as you need to learn it, master it, and move on. Whereas in a traditional school system, almost everyone is moving at the pace of their classroom. Like you're moving at the average pace of the 30 other kids. Yeah. And you can't really speed it up or slow it down unless you have an external tutor, like you mentioned, Zach, that's several hundred dollars an hour. So anyway, that's what I would like to see, more AI teacher influencers and more sort of apps that allow personalized learning speeds. Yeah, I think that that is the right push. And I think what you're saying, and I think we're agreeing, is like the brain, the language models are good enough to teach. Like they really are, right? Like if you want to learn something, ChatGPT is a great place to go learn something. Obviously, make sure everything they're saying is perfect, but it's a great place to learn something. But it's not really an engaging product, especially when you have to learn it and there's stress involved. So I think all of the kind of infrastructure around avatars, around voice, and everything you're describing, Justine, is going to actually unlock a lot around education. I'm sure we'll unlock a bunch of use cases around more engaging or less engaging things that need to be more engaging. Yeah. Right. So I think it's interesting. We're at the moment where I think LLMs are trusted as a learning partner. Now we have to build experiences around them that feel native to how you interact with language models. Yep. And native to the tools that are, one, that we can build with AI today and two, where students are engaged, right? Which is like the Brain Art video is a great example of all of that. Amazing. Well, thank you for educating us all on this, Zach. This was awesome. Yeah, of course. It was great talking to you both.