Edtech Insiders
Edtech Insiders
Week in Edtech 3/19/25: OpenAI’s Writing Model, Gemini Canvas Goes Free, Zoom’s Chain of Draft, China Adds AI to Schools, EdTech Office Closes, and More! Feat. Annie Chechitelli of Turnitin & Sara Mauskopf of Winnie
Join hosts Alex Sarlin and Ben Kornell as they break down a pivotal week of AI announcements, edtech disruption, and education policy shifts.
✨ Episode Highlights:
[00:00:00] China introduces AI education for 6-year-olds, sparking urgency for U.S. to respond
[00:05:23] OpenAI teases a creative writing model as Google and Anthropic push into coding AI
[00:07:59] Gemini Canvas and Deep Research go free, redefining educational productivity tools
[00:11:09] Copyright clash: OpenAI wants to train on protected content, creatives push back
[00:14:02] Gemini’s UI outpaces OpenAI with real classroom use cases
[00:17:54] Chain of Draft from Zoom cuts AI costs by 90% and mimics human note-taking
[00:18:34] Baidu, Alibaba launch emotion-reading and multi-modal AI models in China
[00:20:50] Manus, China’s autonomous AI agent, sparks global interest in multi-agent systems
[00:23:36] U.S. vs. China: centralized AI strategy meets decentralized innovation culture
[00:27:57] U.S. Education Dept. shutters Office of Ed Tech, leaving a national guidance gap
Plus, special guests:
[00:30:11] Annie Chechitelli, CPO at Turnitin, on launching Clarity for ethical student AI use
[01:03:49] Sara Mauskopf, CEO & Co-founder of Winnie, on expanding into K-12 and the rise of school choice
😎 Stay updated with Edtech Insiders!
- Follow our Podcast on:
- Sign up for the Edtech Insiders newsletter.
- Follow Edtech Insiders on LinkedIn!
🎉 Presenting Sponsor:
This season of Edtech Insiders is once again brought to you by Tuck Advisors, the M&A firm for EdTech companies. Run by serial entrepreneurs with over 25 years of experience founding, investing in, and selling companies, Tuck believes you deserve M&A advisors who work as hard as you do.
[00:00:00] Alex Sarlin: I think just the idea of feeling behind is a really good inspiration for sort of racing and really getting hungry for change and actually making big swings. So when you see, like this announcement today, right? It's the Beijing schools are basically introducing AI courses across the curriculum for students as young as six, including.
Fundamentals, chat bot use ethics. They've done it in 184 schools. They're gonna go bigger and bigger. You're like, well, we talk in the US about how AI is gonna change the world and our students. We should make sure that they're ready for it.
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
[00:00:48] Ben Kornell: here at EdTech Insiders. 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.
Hello, EdTech Insider listeners. We are back with another week in EdTech, and my fearless founder, Alex Salin, is back with us. Alex has been on a couple podcasts here, but you are officially back from paternity leave. Congratulations again. Anders Salin, welcome to the world and also welcome EdTech Insider listeners.
We are in the sprint stretch here to A-S-U-G-S-V. We are excited to see you at the happy hour that we always do. If you want a little bit of taste of EdTech insiders beforehand though, we have an AI and educational game development webinar on March 27th, and this is going to be a happening event. Alex, it's an all star lineup and we've had thousands of people who have said they're coming.
I mean, this is, I'm like blown away how much people are into education games and connecting with all these different companies.
[00:02:08] Alex Sarlin: Yeah, it's really interesting. As we were doing our generative AI map of use cases in K 12, one of the use cases we really found intriguing was this idea of multimodal content.
You know, you and I talk about multimodal content and we, we've seen podcasts with Google's audio overview. We've started to see video with YouTube and runway and all these things. But one area that's really nascent but so interesting is the ability to use AI to quickly generate educational games. And that idea sort of spurred me to start to look and pull together some of the people who are doing really interesting work in this.
And we reached out to them and I think, you know, because this is such a new field, we got incredible people. So we have Rebecca Kantar, who's the, the head of learning at Roblox. We have Allison Matthews from Minecraft edu, which has done incredible educational gaming platform. We have Alicia Lee from Rosebud ai, which if you haven't checked out Rosebud ai, it's really interesting.
Basically educational game development. Platform that uses ai. You tell it what kind of game you want, what you wanna teach, and it starts to make games for you. It's pretty amazing. So we brought together all these folks as well as we have a, like 15 speakers or so, people have really responded. We've gotten, it's the highest registered webinar we've ever done at EdTech Insiders, and it's gonna be two hours, basically four panels in a row with all sorts of different people.
We're talking open worlds, we're talking teaching durable skills with gaming. We're talking teachers as game designers. It's a really exciting topic and I think people are intrigued. So I can't wait. It's gonna be great.
[00:03:37] Ben Kornell: It's gonna be awesome. We hope you all be there. Check out our LinkedIn or you can go to our Substack.
Lots of links to sign up there. Also on the pod, we've got an incredible lineup of folks, Michelle Ree, which was already released. Lawton Smith from literal, a longtime friend, Brent Bushnell. We've got Pierre Laman. Rob Barnett from Modern Classroom Projects. I mean the number of folks that are launching new products, sharing new strategies.
It does feel like this 2025 is like a jumpstart for ed tech.
[00:04:10] Alex Sarlin: Yes. I think people are popping off in these different ways. I mean, when I talked to Pierre Lamont, who's the, the head of the Project Management Institute, he is so bullish and excited about what AI can do for project managers, for the education and project managers for spreading the word about their certifications for creating new certifications.
The new technology has created a breath of fresh air for a lot of people, and they're thinking about all sorts of new things. Bran, Bushnell, you just mentioned, he is the founder of Two Bit Circus in la, which is a arcade, basically a giant arcade. Really interesting place. He's the son of Nolan Bush now the founder of Atari.
So he comes from this legendary gaming family, but he also has these educational nonprofit arms where they do things in schools to do all kind of creative gaming and creative, you know, putting materials together. It's really interesting. He's also actually one of the guests at the gaming webinar, and he's really interesting conversation, and he's somebody who's just full of energy, always has new ideas, but AI has taken it to just to another level as well.
He's like, what could be possible? It just blows his mind. He's like, students can create incredibly complex and thorough experiences for each other with the help of ai. So it's a neat moment. So speaking of ai, Ben, do you wanna talk about A-S-U-G-S-V actually for a moment? Speaking of sold out events,
[00:05:23] Ben Kornell: I think we should jump into ai.
I mean, y'all know that our event sold out in. Nine minutes and 20 seconds. So if you're listening to this now, please make sure you sign up for the wait list. We're gonna be lautering off at least a hundred spots. We have a couple sponsors that are rallying to see if we can add maybe 200 more slots for the happy hour.
So stay tuned. We want everybody to be able to come. Last year we had a thousand people. On the registration and wait list, so it, you know, we're trying to make sure that we're as inclusive as possible, so please be patient with us. But as you were saying, there's a lot going on in AI this week. And generally, you know, the staccato beat of AI announcements doesn't penetrate the psyche anymore.
There's just like this drumbeat where it's like almost in a, a wash, but I feel like we are seeing the next episode. People often say we're just in the first inning. I feel like we're in the third inning now, where people are really launching AI with intentionality around large industry use cases and Gemini and their announcement, which you covered as well in our newsletter and on LinkedIn, they are making a big play for coding.
They're making a big play for this text to audio. And I would say anthropic is clearly in the lead in that realm of AI as a coding assistant or even a replacement. And open AI now is teasing a new model that is supposed to have really high capacity at creative writing of all things. Exactly. Which of course is not the industry that everyone is so keen to disrupt.
But it does feel to me like one, you know, I was like gonna sign up for the $200 version of OpenAI just because I'm like, okay, I wanna see this. And then immediately after, there's all this stuff that's essentially as good, that's free, and it does feel like OpenAI is losing a little bit of of its edge.
And you've got these like really, really strong like anthropic, they're B2B is just going so strong. Their sense of use case. And then the Gemini strategy of AI everywhere seems to be working out really well. And I feel like meta is far in the dust. Then you've got open AI who really understands press and how to do press releases, but I'm not sure that they've really gRED how to, no pun intended.
Build something beyond chat GPT. That becomes like an everyday fixture of I. Like coder behavior or user behavior?
[00:07:59] Alex Sarlin: Yeah, it's really interesting. I mean, so we covered a long time ago when OpenAI first created its custom GPT store or whatever you'd call it, catalog and where people could create and share GPTs at the time.
You know, I think the goal was for them to open up user generated content, make a new app store, Google created its gems, which are sort of a similar model where people can create and share applications created by ai. And that had been until this week, only an advanced feature. It was only a paid feature of the gems, but this week they actually moved it into.
They moved it in front of the paywall and now Gemini users can use and share gems even outside of being paid. They also introduced this new feature, which I mean, it really blew my mind when I saw it. I, I tried to cover this on LinkedIn and I tried to convey the excitement I have for IT relationship to education, because the use cases they were showing were educational use cases.
They really make, it makes sense as an educational tool in a lot of different ways. It's basically called Google Canvas, and the idea is it allows people to have a sidebar of a conversation, but while they're having the conversation, it's generating either a document that keeps updating as you do it, but you can also edit the document directly.
So it's sort of like writing in real time, co-writing in real time with an ai. But you can actually get in and, and write. You can change the names, you can change the wording, you can go in and and add paragraphs and it creates a really interesting flow that I think will be interesting, especially for creative writing as well as, you know, all sorts of document creation.
And as you mentioned, Ben, it's also relevant to coding because you can. Build applications, you've basically used common language. You could say, Hey, I want an app that does this. It writes the code of the app for you. And you can then preview and actually use the app inside the Gemini browser and see if it's working and keep tweaking it and adding it and changing it, and then share it.
It's an incredible tool for educators who wanna be able to build their own applications for their students that are custom and bespoke. You can do it without having to know how to code. Now it's incredible for educators who wanna use it to create, you know, lesson plans and other types of academic documents and it's gonna be even more amazing for students.
So I thought it was really exciting and I really encourage people to look at it. And then to your point, you know, open AI's, deep research, Google's deep research, everybody sort of jumped into this like very thorough deep research. At the same time. The creative writing announcement from OpenAI, I don't know where that came from, but.
I'm intrigued by it. I mean, I just think it's interesting and I can imagine there, there's probably an education use case of all things for them to focus on. That seems like, if I would imagine they're stacked list of what areas they want their AI models to get better and better at. I would've thought the creative writing would be way down at the bottom, but for some reason it is not, and that's intriguing.
We also saw, by the way, in a relevant story this week, we saw a bunch of celebrities basically sue, try to complain against OpenAI and Google saying that, Hey, you're basically eating our content. You're taking all these things that we do. You're imitating us, you're taking our copyrighted material. Like this is not okay.
And open AI going the other way. They're saying, Hey, AI models should be able to train on copyrighted material and asking the government to do that. So there's increasingly this tension between creatives, real life, human creatives, and these models, which we've seen for a while, but it's pitching up now.
It's a wacky moment.
[00:11:09] Ben Kornell: Yeah, it does seem to me like the open AI going with a creative writing pitch is off tone from a business standpoint of like where they need to be going. And by the way, English major history and nurture major over here. I love me some creative writing, so don't get me wrong. And. If my next fiction novel is just reading what Chet's writing for me, I'm thrilled.
But I will just say we talked with the notebook, LM founder at the Google AI Summit, and one of the things that was my aha moment in that was really user interface as the innovation. And so what you have is like an AI system and then there's something in between that and the human that allows for the interactivity.
Turns out generally it's a computer screen with, you know, computer keyboard and all of this, but notebook lms, part of their innovation was the triptych where they kind of have a before middle and after where you're able to upload your things. This is another move where Gemini is saying, how do I think about your physical.
Space with ai and where can you do your like prompting and your dialogue over here and where can you see the outputs over here? And you know, basically as you described Canvas, that's part of the nuance. And I think chat, GPT has in some ways been optimized for a mobile phone screen. And you see that anthropic clouds on it is very much like built for a broader computer screen.
And what you're gonna imagine for like an inch team person, they may have a double screen and it may be spread across. And by the way, maybe computer interfaces aren't even the right interface. You know, humans might look back and be like. Why were they looking at screens all the time? This is totally the wrong way to interact with ai.
I mean, I was at the computer history museum with my kids in Mountain View and they had these like massive, massive computers and you're feeding in these like hard readers, you know, with holes punched in them and there's like not really even a screen and the kids are like. How do they even know what the computer's doing?
They can't look at a screen. There's no keyboard. You know, what we're seeing here is that the technology itself is reinventing the mediums that are going to work to take advantage of it. And this is where, you know, I think voice and texts and video and all of that stuff can also be really, really interesting.
But you know, who knew that Gemini would be doing laughs around some of the other people with ui? Given that Google's UI in the past has been like one of the, it's so simplistic that people never really have given them credit as a a UI company, but maybe it's that simplicity and practical use case that's actually.
Being an advantage for them.
[00:14:02] Alex Sarlin: Yeah. And well, they've also had a long time to get it right with a ridiculous numbers of users using it all the time, which I think is really interesting. I mean it, they work in the billions of users and yes, OpenAI is the fastest growing company ever. Chat GT is the fastest growing tech product of all time.
So they've certainly built a very big user base very quickly. Google has had a very big user base for a decade and they literally are counting the billions, you know how many people are using docs and maps and things like that. So they have to have a very, very clean ui. You know, we've talked about this for a while.
I do think that the form factor, like you're saying, whether you're doing it on a phone, whether you're doing it on a laptop, whether you're doing it at an engineering station with two or three screens across, or whether you're doing it through an Apple Vision Pro, where you can literally look around and that's how you control your mouse, quote unquote, is gonna be a really big aspect in what AI looks like in the future.
It will change how we interact with it. And of course that also. Comes down to education, right? I mean, when we think about how students interact, we talked to the founder of Jotted recently, and I talked to Levi Bena from Merlin Mind recently as well. And both of them are really thinking about the classroom environment and the form factor of the classroom.
And what does a teacher need to be able to do? They're standing in front of a set of kids. Every kid often has a one-on-one device that they can look at that can either be helpful or distracting. They also may have a phone that may be helpful or distracting at the same time, the teacher has to organize things.
They have to make groups, they have to get everybody's attention, and they're both thinking in different ways about how you might use AI to create, you know, more power basically in the classroom for a teacher. Not power in the sense of top-down control, but power to be able to do things in real time.
Like take the pulse check of the class or pop things up on people's screens or set up groups and say, you three are in this group, boom, suddenly your screens are all aligned. You three are in that group. Boom, your screens are all aligned. Like it's a really. Cool vision of what tech could look like, that I think takes us beyond traditional consumer tech.
It's really exciting. One other thing that jumped out to me, it's a little bit of a different piece, came from a different direction than, than the traditional AI companies we talk about here. This is actually a finding from Zoom. It really caught my eye because it was something very, very interesting.
Basically, there's this AI model called chain of thought performance where they basically require an AI to go through its logical steps. It would be step by step, and by doing that. It comes up with better conclusions. It avoids going off track. It sort of has a better performance. And they came up with this new idea called chain of draft.
And the idea behind it is rather than having to list out everything you're thinking, it tries to replicate what people do. What is, it draws inspiration from how people go from thing to thing, which is that they take notes, they highlight the key things that matter. They say, oh, okay. They underline, right?
They don't write down every word in a transcript of a call and then go to the next call and read every word of the transcript. They take away a few key ideas, and that's what this is too. It's saying if you can train an AI to take a few key ideas out of each piece of its thought process, it can get incredibly good results with much less time and much less compute, which is really exciting, especially for enterprise.
So this is a totally new idea, and I don't know if it's gonna take off or not, but this is this idea of chain of draft. There's also this whole new. Idea of a model. Speaking of meta called the LCMI heard about this week. Again, I don't know how big this is gonna get, but it's this idea of large concept model where instead of pure language, you're actually trying to build a model that understands conceptually how things are connected.
And that is a sort of a different way to think. Maybe not constrained by language and maybe more conducive to the omni modal future. We're sort of thinking about, I can't talk about it that much 'cause I don't know that much about it. But if you're interested Google LCM, and I think this is came outta metal researchers, but it could be a really new way to do it.
So, I mean, you're talking about the third inning and I think the next few innings might be actually step changes even in the underlying architecture of some of these models.
[00:17:54] Ben Kornell: Yeah. As you're talking through all of that, then I'm thinking about societal components of like, okay, how do we as a society. Take advantage of it, but also reconcile that, whether it's AI mimicking our chain of draft or whether it is we're moving to these deeper thinking models.
One glimpse into that future is really coming from China. I know, Alex, you curated some of the new from China. I mean, what stood out from you? We're really in this market-based like Wild West over here, and it seems like China is taking a little bit of a different tact and it is actually quite compelling.
[00:18:34] Alex Sarlin: So the way I would frame it, I'm not a, you know, an expert on Chinese politics or anything, but the way I would frame it from what, how I'm sort of seeing it from the outside is that China, like America has a few enormous tech companies that really. Dominate the ecosystem. And just like in America, they have really advanced AI models.
So we saw this week two big announcements from big tech companies in China, and I don't think we should underestimate the power of these. We saw Baidu launch new AI models, huge company, huge AI expertise. And we saw Alibaba release a new AI model specifically to try to read emotions from videos and sort of be more advanced computer vision and visual model.
So we should pay attention to those because these are huge companies. They think really hard about how to run the space. And then on the flip side, you see these startups. And the startups, some of them are doing things that are taking the AI world by storm and really surprising them. So we saw the deep seek moment a few months ago.
I'm not as. Deep as some others are. But deeps seek has been growing in China. It's now being incorporated into lots of different spaces in Chinese ai. And the hot startup this week is this company called Manis. I think I'm pronouncing that right. It doesn't sound Chinese, but Manis, which is basically an agentic ai and it's a Chinese autonomous agent that's designed to basically go through lots of different systems, make choices, and put together, you know, final results.
So it's a multi-agent architecture they call it, or it's built on a multi-agent architecture, which means it can sort of. Basically do a whole job, which even though it means a complex workflow that jumps from program to program or task to task, it can do it all together. So we saw MIT Technology review put out an article this week where they tried out Mattis and they said it was pretty good and it couldn't do everything they asked it to do.
But the things that it could do, it would do very thoroughly and with like multi-agent. So people don't, aren't sure how big a deal it is, but just the idea of Manness and deep seek are these startups and then Alibaba and Baidu are these giants. And just like we see here as sort of things, the open AI startup model versus the Google and Meta and Apple, you know, incumbents.
I think something similar is happening there. There's also something happening in the schools. But before we get into that, what do you think about the Chinese AI ecosystem? Do you think we should feel inspired by it? Threatened by it, or you know, what can we learn from it?
[00:20:50] Ben Kornell: Well, I actually want to talk about the schools 'cause I think it's deeply connected.
Go for, on the school side, the Ministry of Education is making a commitment that with ages as young as six, that AI education will be infused in the curriculum. And if you look at the last 20 years, China has been pumping out. Engineers and technology leaders and so on, and the knock has always been the entrepreneurial energy.
The creativity in the US still gives an advantage in our system, but as you start seeing ai, that one, some of the AI I. Is about creativity. So ideation, exploration, there's a way in which you can bolster one's creativity. And two, when you're in a more planned economy where jobs are more scripted and actions are more rote, they are better lined up for AI displacement.
Whereas if like I run the city of San Francisco, how I collect garbage and how I run my city is totally different than any other city. I mean, it's not a for-profit, it's not a startup, but it's entrepreneurial right there. It's like, okay, we have 70 cities, all of which are larger than San Francisco and all run exactly the same way.
And now if you have an agen AI that you could put in there, think of the scalability and efficacy. And then last I would say. This is all about skating to where the puck is going. You could snapshot where China is today and say, okay, for some things they're ahead, some things they're not. But what is the secret for AI leaping forward?
It's the number of users and the data. They've got a billion users that they're able to, you know, get using their AI and learn from it rapidly, and they're starting at six years old where all of those folks are gonna be prepared for an AI future in an intentional way. And of course, it's not gonna be all right at the very beginning.
But then you contrast it with what we're essentially not doing here in the us. I think that this is like the industrial revolution where there were new technologies that created winners and losers or accelerants for different countries and cultures. This is such a technology, and whether it is full replacement or whether it's human augmented, if you augment a billion people, that's pretty damn good.
And if you do it in a way that's systematic and thorough, that's pretty damn good. So we're gonna see how well our entrepreneurial, capitalistic, decentralized system works in a world where large data, large implementation. Large rollouts offer sufficient scale advantages?
[00:23:36] Alex Sarlin: Yeah, I mean, I think that's such a good analysis.
Like it makes me think of how we talk in America about our Sputnik moment, right? The moment when Russia launched its Sputnik satellite and the US was like, oh man, we're behind. We have to catch up fast. And it was this huge resurgence in science education and university funding and basic research. It really was like, uh oh, we've gotta catch up.
We gotta go. And I think some of the countries that have been mired in political oddness for a long time, China is at the top of that list, but there are many others. I mean, in the fifties, China was having a cultural revolution. It was like going in this bizarre direction. And then a lot of the Asian tigers, they looked up and said.
Wait a second. The world is changing. We have all of these people, we have all of this potential. How do we leapfrog? How do we move really fast? And in a somewhat, you know, centralized system like the CCP, you can actually just decide. And so they've decided some big things over the last few decades, especially around education that have made a big difference for their GDP growth and for sort of the way China has moved in the world.
And it's not always consistent and it's not, certainly not always fair or doesn't always work out well. It's not to try to glamorize this, but I think just the idea of feeling behind. Is a really good inspiration for sort of racing and really getting hungry for change and actually making big swings. So when you see, like this announcement today, right?
It's the Beijing schools are basically introducing AI courses across the curriculum for students as young as six, including fundamentals, chatbot use ethics. They've done it in 184 schools. They're gonna go bigger and bigger. You're like, well, we talk in the US about how AI is gonna change the world and our students.
We should make sure that they're ready for it. But frankly, I don't think we know what that means. I mean, I think that's this incredibly hand wavy thing and they're like, we don't necessarily know what that means either, but we're gonna try. Instead of doing paper after paper and thinking about it and trying to like convene committees for the next 10 years before you do anything, they're like, let's do it now and let's go all the way down to young elementary school students.
It's admirable in certain ways. I.
[00:25:43] Ben Kornell: The contrast is so profound because here we are doing culture wars in schools. Yeah. And we, we are not aligned. And there's this big moment here, an inflection point and our listeners will say, okay, inflection point are you guys being overly dramatic, but there's a way in which like the next two decades, let's just not, we don't have to be so myopic the next two decades we'll really launch the next wave of technology.
And technology from like agrarian to industrial to today has been the lever of acceleration. If you look at the way that China invests in education and believes in their future generation as the key to continuing to rise as their country, that's a driver. The second driver we also have to remember is that they have a social contract.
The social contract is you give up a little bit of your freedom and we will make your life better decade over decade going forward. And they actually have a legitimacy urgency because they were starting to slow down. Their growth was slowing down Covid. We look at it over here as if like, okay, they navigated covid the way they did.
It's controversial there, like should they have shut everything down? Should they have done what they did? So there's a government mission, which is, okay, how are we gonna use this technology to leap forward because that's important for us to maintain. I. Our authority in the country.
[00:27:14] Alex Sarlin: They wanna inherit the world.
I mean, they wanna be the main power in the world. I mean, very much so.
[00:27:20] Ben Kornell: And I will just also say there's really smart people in these positions of government that are making complex decisions. And this is not a group that maybe in the Mao Cultural Revolution era, maybe the education levels or maybe the decision making, they were kind of emergent.
They have top scientists in the world. They have top thinkers. They are now playing with a team players across lots of key positions. So, oh yeah. I feel like it's super intimidating. And then meanwhile, we've got our own geopolitical stuff, which we won't get into.
[00:27:57] Alex Sarlin: Well, we, let's get into it a little bit because there's some very relevant things this week.
We don't have to go deep, but I mean, we saw the Office of Education Technology this week be dismantled. It was the federal office, the Office
[00:28:09] Ben Kornell: of
[00:28:09] Alex Sarlin: EdTech.
[00:28:09] Ben Kornell: I mean, this is a blow to our space, you know? Yeah. And EdTech had been one of the areas where all parties said, we've gotta figure out EdTech. I feel like there's one way in which this is just part of a mass cost cutting, but there's another way in which I think things have soured on EdTech, the public support for EdTech.
There's a lot of pushback, and I think it's mainly with kids and screen time. I don't think in higher ed or career. That it's getting the kind of pushback, but I'm just seeing a lot of people trying to move away from screens and back to paper and back to notebooks, just almost as if technology bad and yeah.
So it will be interesting to see how all of this is absorbed because I know a lot of the funding that people have been getting with SBIR grants is gonna go to the NSF. But where does Office of EdTech go? Probably nowhere disappears.
[00:29:07] Alex Sarlin: Nowhere. I'm sure there's not gonna go anywhere. Okay, that's it from us today.
Now let's turn it over to our guests. Annie Elli is the Chief Product Officer at Turnitin, where she leads the development of applications focused on academic integrity, grading feedback, and assessment. With over two decades of experience in EdTech, Annie has been at the forefront of expanding access to education and supporting educators and students.
Before Turnitin, she spent over five years at Amazon leading Kindle content for school, work, and government, and launching the AWS EdTech Growth Advisory team. Annie's EdTech journey began at Wiba, where she developed a live collaboration platform that was later acquired by Blackboard, where she led the transition of Blackboard learn to the cloud.
She holds a BS from Columbia University and an MBA and MS from Claremont Graduate University and lives near Indianapolis, Indiana with her family. She's an avid tennis player and she competes in USTA Leagues in both Indianapolis and Seattle. I. Elli, welcome to EdTech Insiders.
[00:30:11] Annie Chechitelli: Thank you. It's great to be here.
[00:30:12] Alex Sarlin: It's great to chat with you. So we've gotten a chance to chat casually in a couple of conferences, and you and I were on a panel together. I was moderating a panel that you were on a couple of years ago at A-S-U-G-S-V. You always have incredibly interesting insights. There is news from Turnitin. So we wanted to bring you on the podcast and talk about Turn it in.
But before we do, can you just quickly give a little bit of an overview of what Turnitin is for people who might not know and what brought you there in the first place?
[00:30:39] Annie Chechitelli: Yeah, so Turnin is a leader in academic integrity solutions for both higher education as well as K 12. Globally. The company has been around for 25.
If you think you know your way, way back machine, the internet came out and people were afraid that students were just gonna copy from the internet and never do their work and not actually develop skills. Sound familiar?
[00:31:05] Alex Sarlin: Yep.
[00:31:06] Annie Chechitelli: Obviously it's a very different power, but the allegory is there and some teachers came up with a solution that would look at the internet and compare it to student submission to help teachers identify if students were either maybe not citing properly or not really understanding how to cite the internet or taking shortcuts such as just your standard copy and paste.
I mean, when I went to school, it was just the encyclopedia. That'll a little bit but started back then. As a solution to add elements around, like how do we give feedback to students? How do you incorporate that into a submission? Let's not just at the along the way, contract and essays became also an I to education, and so put out a solution that would help identify when a student may have turned in a paper written.
20 19, 20 20. Turnin did some additional acquisitions, both in the integrity but also kind of in the assessment space. Because when you think about it, an essay in a lot of ways is an assessment. So what are other ways to think about that, both from a AI perspective, but also from an integrity perspective.
And so turn and purchase a company called ExamSoft. Not a lot of people know that ExamSoft specializes in secure exams. So the bar exam or a nursing exam, or a medical exam, any of those high stakes exams, as well as figuring out how to qualify the questions is like when you think about those types of exams, it's not like a teacher just comes up with a question in a more formative way.
Those questions have been rigorously tested for what level, what's the success rate, and all of those, all that telemetry in that system, as well as another product called Grade Scope, which is all about. How do you bring that paper world to the digital world in the best way to maximize efficiency for instructors and also just make it digital?
And so thinking about how that, anyway, that product has been interesting because I'm sure we'll talk about, you know, a lot of teachers are handling the response to LLMs very differently. Some are like, we're just going, I'm just going full Blue book. I'm just gonna go in class. And then thinking about, okay, well how can we help that in addition to the work that we do?
And so it's been interesting. I would say turn it in very global. People don't realize that as well. Half of our customers are outside of the US and so it's also been interesting to look at this from a US perspective, but also what's happening in Australia and the uk, which, you know, academic integrity in some ways is treated at a lot more scrutiny than in the us.
And so it's been interesting to see the different markets.
[00:34:08] Alex Sarlin: Oh yeah, as you alluded to, the internet was one wave in, you know, academic integrity concerns. Then essay mills and you know, the idea of contracting and buying assignments, which also was sort of facilitated by the internet. And now we have this AI moment and everything's just up in the air.
Again, you know, you've been right in the forefront talking to people for a while now about their academic integrity concerns, and as a result you have launched a new product called Turnitin Clarity that pieces some of these pieces together. Tell us a little bit about what you've seen out in the market about some of these different reactions you mentioned to AI and academic integrity and what inspired this particular product launch.
[00:34:48] Annie Chechitelli: Yeah, so it has been quite a journey. So we actually launched our AI indicator two years ago, which was the first thing, which is like, what are we seeing? Just like, help me figure out what's happening. And so we continued to offer that to our customers as well. And we started talking to customers kind of around areas of kind of why they're still finding this hard.
They're kind of in this middle of like, well, I need more data points. Like I feel like this is okay. I have some data points. And many times I have found, generally those conversations between the teacher and student tend to be easier in the high school setting than the college setting, because in the high school setting, they know their student more, they see them more frequently, they have more pieces of work.
It's just easier for them to manage that than in college. And so looking at. Kind of the evolution. And so that was part of it was I need more data points to verify the authenticity of the work, love the AI indicator, but kind of want more so I can triangulate more. And so that was the first bit of information.
And it was interesting that the past year and how the sentiment has changed, you know, you have this, I would say first we got through fear pretty quickly, and then we got to like, oh, it's, everyone has to embrace it. It's amazing. It's the future. And I don't disagree, but that's easy to say. That feels like platitudes at this point of like, okay, well then how do we do that?
Like what is a responsible way? Like it's easy to say that, but what are the actions we're taking both as an industry. And as a community to be able to do that. And I became frustrated, quite honestly, with just hearing a lot and not figuring out or not seeing what is that gonna look like. And you and I spoke a year ago and I told you I was frustrated with the lack of research.
Like what are the elements of writing that need to stay like? And I was saying yesterday in my discussion at South by Southwest edu of like, I want a matrix. I want somebody smarter than me who's actually educated on, okay, on one axis it's a grid of all the ways writing is used, it's used to communicate, it's used to persuade, it's used to evaluate, it's used to all of those ways, right?
And then on the other axis is kind of what are the elements of that piece of writing and what can be outsourced and what should not be outsourced. So we can really scientifically go about, okay, what do we keep, what can we outsource to ai? And that work hasn't happened. And so I kind of felt like the industry.
I'm not saying no one's doing work. I'm sure there, there are plenty. But this moment of frustration, I see a lot of administrators not wanting to get too far involved because they feel like it's a very contentious situation. They don't know the right, there's no right answer, right? But the teachers are like, I need help.
I'm here every day teaching students and you're not helping me with this. You're just telling me, make sure my students learn, but embrace the future. Like what's the guide map for that? And so they were really frustrated. And then the last thing we saw this year was really this frustration of students.
That was kind of the final tipping point for me. And, and it even started, you know, a year ago we'd hear when I was with you at a SU gsb, at the pre-show, when the students were like, tell me how much is too much? Like, turn it in. You must know. I'm like, I don't know. Like it depends on the assignment, your level, your discipline.
Like they were just really. Struggl They at the same time, we started to get more statistics.
Harming their critical thinking skills. And I also, you know, I'm worried about, and I have two kids in college, the job market in the next, I think three years for entry level or first out of college, Jobs's going to be harder. Like they are competing with ai, there's no doubt about it. And so how do we help them understand what is above and beyond AI that they're bringing to the table and how to lead into that.
So that was kind of like the frustration that was in the setting that's happening around. And students also just saying, I wanna learn how to use it. Educators saying, I don't really know. Needing a tool that provided a little bit more insight. And so we thought up clarity, and it's changed. We've done a lot of research the past year directly with educators and students throughout the process learned so much.
And I'm sure even when we launched it kind of in the summertime. It's going to iterate a lot just 'cause it's new. And there are some elements that we know, like we can have a best guess on this, but we're gonna see how the student uses it, how the teacher looks at it, and then we can tweak. But overall, you know, there's two components of turning clarity.
There's this element of integrity, which is what are the elements of the student writing. So you get the basic, how long did it take, how much is copy and paste? Really basic statistics to help the teacher kind of validate. And then in that is integrated with all of our other systems. So if something is copied and paste, they will see where it was from.
Or if it has, and AI exhausts on it, we'll be able to look at that. The second part is that pulls it together is, okay, this is a space where students can use ai. The teacher. And so we have your basic open prompts, and then we also have education specific prompts, like, check my citations, compare this to my rubric.
And that's the part where we're like, we've done a lot of research, but I don't think we know all the prompts and like, we don't even know if we're gonna keep it. Open forever. Like I think that how LLM should be used for education purposes is still new. And so if we put this out there with the guardrails of the teachers kind of watching it and being able to see a high level report if they want, we'll get a sense of that.
And so that's what turn in clarity is that combination of integrity, the process that the student took, but also that guided AI in a really safe environment so students never have to worry like, did I use it in a way that it will be okay, like you're doing it with the teacher can see it. So we can all have an open dialogue and the teacher can start looking at results.
Like, okay, well this assignment, this student used it. These are their grades from the test. Are they evolving? Are they learning? And kind of get a sense for figuring out how to maybe make that quantitative but getting closer. Because right now, teachers, and even in any like B2C model, teachers can't see how students are using it.
They can say how they use it. Maybe some really good teachers are doing things like, submit your prompt to me. But that's different than like see the whole prompt history. How did it come together? How did you repro, how did this, which is very different. So that is turning clarity.
[00:41:34] Alex Sarlin: The words that jump to mind as I hear you describe it is transparent ai, right?
It's transparent to both sides so that it's not, you know, the student skulking off to chat GBT or to one of the tools and not then not knowing how much they can bring back in, whether they're bending the rules because they don't know what the rules are and it's transparent to the teacher. They can actually see, they know what access to LLMs and AI are baked into the assignment, and I imagine they probably have some control over them as well.
I.
[00:42:00] Annie Chechitelli: Exactly. So right now it's, we have on or off on assignment basis, and then we're gonna have levels. So I want just these prompts, like some basic prompts, or I just want like outlining or check my, you know, sentence structure. So, yeah, sorry to interrupt, but got me so excited. No, it's great. So those assignment settings are a key part of it, right?
In an environment we're trying to learn.
[00:42:22] Alex Sarlin: You're creating the rules of the road as they say. You know, when they talk about creating platforms with two-sided marketplaces, the platform creates the rules of engagement. And that's what I'm hearing you say here. You know, hey, citation, checking rubric, checking sentence structure, this is the different paradigms of AI that you can use in the context of an assignment.
And then you sort of allow both sides to use them and you allow the educator to control what's turned on and off, which makes a lot of sense. And then of course you bake in the integrity side. So if somebody does try to bring in third party LLM work, it ideally would be noticed by the system and by the teacher.
And you know, it reminds me, uh, Ethan Molik, as you know, is, has written so much about AI and how his perspective is it definitely should be used in the classroom and it's hard to detect. And his real approaches included in the assignments make it that yes, you assume they're gonna use AI and they should show you and explain how they're gonna use ai.
And I think that's baked right into your approach, which is really interesting.
[00:43:16] Annie Chechitelli: We're excited. We're still getting feedback. It's gonna be hard for us to like cut a date of like, oh yeah, we have enough feedback and then we're on a biweekly basis with it. But it's been great. I've. About students that I never would've expected.
I'll give you an example. Student anxiety is a thing. We all know it's a thing, right? I have three of them myself. It's a thing. So it was fascinating for us because when we were working on clarity, I think we intended to give, at the beginning, we intended to give the teacher access as throughout the whole thing.
Like we didn't really think about. Phases in terms of comfort level of students and students were like, no, I'm totally fine sharing my process with the teacher, but I don't want them to see it when it's still being done. I don't want them to see it early when it's not good, and then have a bias in mind when it comes to grade it.
And like we would've never thought about that. And so we did change it. The teacher can't peak. It's not until the student submits it. It's not until the student says it's ready. The deadline hits in the lms, it'll automatically submit. We built those fun features in there too, just to make it easy, but that the teacher can't see it until the student says, I'm done, which we wouldn't have never known.
And it was just through all the interviews that we saw students just be like, that little thing means a.
The difficult part for the teachers is the user experience. You can imagine how much data we're tracking that we have around like think about when you write like how many drafts you have. And this is funny, we actually started so naive, we started with like more of a textual version history and you could collapse it and like what are the events that will trigger a hierarchical version change?
We so conversations and show. Like, don't make it textual. That's not gonna help them. And it was so much cognitive overload. And then we also hear like, because we can do it, because we do it with our other products. Many teachers just wanna make sure everything's okay. They don't wanna go through all this.
Some do, especially if they're teaching writing, right? And so being able to up level that information in a way that the teacher or the instructor knows like, oh, I should spend time on this. So every like a paste event for instance, gets like a flag or just an indicator so they can just go right to that point and see it.
None of this is, oh, that's not super earth shattering, but to put it in one place in their existing workflow is new. And so it's also integrated with the LMS. So they can say, this is a new assignment. It's not like it's an assignment outside. And so we really wanna make it as easy for the instructors as possible.
But that's hard because it's a lot of data.
[00:46:01] Alex Sarlin: Yeah. It's interesting when you talk about watching the video and, and having flags that you pinpoint moments to look at. There's this really interesting mixed model that you're going for here, which I, I respect. I think it's really intriguing of like the teacher is both being the security guard, right?
They're the ones being like, oh, something happened here. I gotta go look at it and make sure that nothing, you know, unto happened and they didn't do anything wrong. And the teacher, especially as you mentioned, like the writing teacher, where you can actually watch a student's whole process and their drafts and how they use AI and what they outsource to use your word earlier, you know, what do they outsource to AI and what are they coming up with themselves?
Those are like such different hats. And I can imagine that that's something teachers really struggle with now is sort of when to put on that security cap versus the supportive teacher hat where they're trying to really build their students up. I'm curious just if that resonates with anything you've heard in the field.
'cause those, those feel like such different roles to play as an educator.
[00:46:53] Annie Chechitelli: Absolutely. No, I, it's hard to be an educator right now for sure. And it was interesting, I feel that the integrity piece is a check mark for them. Mm. But it's one that's difficult. That's kind of like you think about the hierarchy of needs.
They need to make sure the student did the work so they can spend then the time on it to provide that really useful feedback to make them better. And so they actually really wanna spend the least amount of time on the integrity part. And that's why when we did the research, once we explain kind of what we're adding in there and they're just used to turn it in.
So like, oh, I see how this is all gonna come. Okay, I see how that works. They said, well now that we have this, I want you to do more formative. So that was a base of feedback. And so obviously we added the ai, but we have a whole backlog of, okay, I want multi-part assignments. Let's, we've been hearing that for years.
I want this a moment in time where they do this part and then it connects to this part and then it scaffolds to that group work. We get a lot of group work and it was fascinating. I heard some stories of stories. I'm sure you've probably heard some of friends whose kids were assigned in a group.
Somebody in the group used ai. It like was a big problem for the group because they didn't know. And ugh, group work in general. I'm one of those people who I totally see the value of it. 'cause I work in teams and my team works in groups, but I also as a recovering good student that to be the one who always did everyone's work, that wasn't great either, depending on the group.
And so one of the things where it's in research and development now is showing the teacher actually who did what in the group work, because we can do that now.
[00:48:31] Alex Sarlin: Yep. That's this classic problem in economics and everything, but the free rider piece in group work, there's all this interesting research about that.
And you're right, it's never been possible before, so that's really interesting that it's auditable and transparent and that actually has this added benefit.
[00:48:46] Annie Chechitelli: Yeah, so that'll be part, so the multi-part and group work, those are in like research right now, but we need the foundation as well. This foundation will allow us to move into that.
But anyway, it's interesting you asked about that. Once they have the integrity stuff. But like this can be so much more powerful. And so most of the requests are around formative tools.
[00:49:06] Alex Sarlin: Yeah, no, that makes sense. And I think that's exactly how it should be, right? The integrity is a check mark. I mean, sometimes, you know, I've been looking online because I've been researching this and you see teachers who have all these like tricks and techniques to spot integrity issues that people are like embedding Trojan horses into their assignments and then if they show up in the end it means the ai, like there is this sort of a cottage industry among teachers of trying to sort of spot the cheating.
But I think that's few and far between. It makes so much more sense that teachers don't, they, of course, it's a foundational, like you say, bottom of the hierarchy of needs. You have to have done it yourself. And if you do and you did it with integrity, then I can go in and actually really make sense of it and give you meaningful feedback and look at your process and do multi-part and, and really sort of make sense of it.
I think that's a healthy attitude for teachers, but I'm sure it's really tricky right now.
[00:49:54] Annie Chechitelli: It's so tricky.
[00:49:55] Alex Sarlin: Yeah.
[00:49:56] Annie Chechitelli: My husband's a high school English teacher who uses Turn it. He sent me one of the first ones. It was a screenshot. It said a hundred percent ai. He is like, well, what's this? Because he hadn't used it before.
I, well, the student that's, they used ai and it's been nice though, because for him, his students are, he's a very understanding, empathetic teacher, which I'm a big fan of. Like, let's have the conversation, let them learn from their mistake. You know, they're still young enough, let them redo the assignment.
Maybe they can't get an A, whatever that looks like, and I want that. Every teacher has, it's a very personal thing, like integrity in the classroom. And that's the other reason we design this, is everyone wants to do it differently. I'm not gonna impose that upon an institution. Like they can't, like, this teacher wants a detector, this teacher wants this, this teacher wants that.
Like whatever it is. But every conversation he has had with the student, they've always fessed up in some way pretty quickly. I think they're just like, oh, my uncle helped me. And he like did, he was like, yeah, okay, well why don't we try it again? Because this looks nothing like the draft.
[00:50:53] Alex Sarlin: My uncle Claude.
[00:50:55] Annie Chechitelli: Yeah. Yeah. That's. I, I'm gonna steal that from you, my Uncle Claude.
[00:51:01] Alex Sarlin: Exactly. So families are talking about AI as well. It's when you mentioned this idea of students wanna know the rules, they wanna know what they are allowed to use, and the fact that it differs between teachers or differs between subjects probably is pretty confusing to them.
I'm curious how you think that's gonna evolve over time. Like as you mentioned for educators, integrity is a very personal thing. They wanna set their own rules, but for a student who's in six different classes with six different teachers. I guess this is part of what Clarity is meant to address, right?
How would they know what they are allowed or disallowed to use in any given assignment or moment? It feels like that's probably apart from what you're addressing.
[00:51:38] Annie Chechitelli: Yeah. I dunno if we're solving it and I think like I go back to the, once again, the internet way back machine. I remember students being like, but this course in the L LMS is set up different than that course.
Like they figured it out over time. I mean that's a very different situation, but like it does create differences that the student has to adjust to for sure. And for me, I see it more at the level, like I understand teaching MBA Wharton, if those people do not have critical thinking skills, by the time they're there, we are in trouble.
So use away, but like senior high school, freshmen, even sophomore in college, I think we're still trying to make sure that they're developing that so they don't come out in a way that they can't add value above what AI could do.
[00:52:22] Alex Sarlin: That makes sense. It reminds me of the conversation and the metaphor that we've been using, I think as a industry for a couple of years now, which is the calculator, right?
Yeah. When the calculator first came out, it was considered an integrity risk, and then there was sort of over time this decision of at what level of math do you need to be at to be able to have a graphing calculator in your hand? And we can assume you can do the multiplication and do the fractions, but to actually do X level of math, you're allowed to have a calculator.
And I feel like it's now writing is going through that same process. And to your point, we don't have that matrix. We don't yet know what you have to prove that you can do in terms of critical thinking or persuasion before you can then get an AI assistant who you can just say, Hey, make this more persuasive and it'll do it for you.
But you know, you have to show, educators are responsible and students are responsible for knowing that they can get those skills themselves without it being completely outsourced. It's a really tricky moment. So one thing that as somebody who's been in the classroom doing all this user research.
Rolling out products and of course already having one of the most widely used globally platforms for integrity. Are there things that we in the ed tech industry who maybe spend some time in classrooms, but probably not as much as we could or we should? Are there things that you would say are really surprising to you that you feel like the industry should know about integrity or about AI usage?
You've mentioned a couple, like students themselves are afraid of losing their critical thinking skills or that group work and AI are a funny combination. I'm curious if there are other things that have sort of surprised you that you'd like to share with our audience.
[00:53:52] Annie Chechitelli: I think one of the surprising ones, and I, it's not that surprising, but the rates of which the traditional.
Misuse happens. So students just doing copy paste, that hasn't changed. Hmm. Like that's still a good fallback for a lot of students. We look across all of our products at kind of where that score sits, just to understand it, but it hasn't changed. I think the surprising thing for me is the links at which students will go to not do work.
Like I made a montage for our company kickoff at the beginning of this year of all the students. One of was like, I've been paraphrasing for three hours and. Still coming up as AI and I'm like, I dunno how to help you buddy. Like just do the work. But then even like, I think the surprising thing for me is the secondary economy that comes up on these things, like the Humanizer tools, like 1299 a month, they can take it from like Uncle Clade or Chad PT and try to get around a detector.
And there's been so much that have come up and they're not cheap. My favorite, I told, the only thing that makes me feel good about that is like we're all, turn it in is always listed as like you have to have the premium subscription, which at least I get that shout out by the cheaters. But I don't understand, I mean that secondary market is surprising and how fast it came up and how many there are and I can't tell if they're all just like the same thing with different storefronts.
Like I don't really understand, but I. We continue to work to, as these things pop up and students brag about them in public places,
[00:55:32] Alex Sarlin: right?
[00:55:32] Annie Chechitelli: We do incorporate them, and at the end of the day, a lot of this stuff is just statistical math, and so we'll keep on those trends as they happen, as more companies pop up in this economy.
[00:55:44] Alex Sarlin: Yeah, I have been very surprised at that too, and in doing all this research about different AI ed tech companies, some of the fastest growing and most popular ones are these homework help slash snap and solve, humanizing. For those who don't know that word, it's like you can imagine, right? It's taking an AI generated response and trying to make it look more human by adding errors or different kind of grammar things so that it will pass AI detectors.
It's like such a warped circle that we're going in. You get this polished essay from ai, then you humanize it back to look somewhat like you, but obviously you want it to be less human than yourself or else it is not gonna get the grade you want anyway. They're really popular. They've grown really fast.
I don't understand that market either. Some of them, I've tried to actually find the founders of them and you can't find them. You can't actually identify online. You can't even find who the people are. Sometimes there's students in overseas universities who did it all themselves. Sometimes it's just gone.
There's literally, you can't find who's behind these tools that have many, many thousands of users. So whenever I talk or write about this, I always find myself using the word arms race because it just feels like you said, turn it in. Opposites, statistical math and increases its ability to detect, and then the tools try to increase their ability to humanize and surpass it.
My question for you, and I know that you really care about this, integrity is obviously a core part of what Turnitin does, but is there going to be a moment where we can sort of. Move past the arms race feel of this, where it's like teachers are trying to get techniques to spot students, and students are trying to get techniques, even if it takes hours to be able to find a way around having to do it completely on their own.
Do you feel like this kind of sanctioned AI where there's a slider about what you can do and it's transparent and it's auditable, do you feel like that's gonna be the way that we sort of get past the arms Royce mentality and get to a place where everybody's sort of more on the same page?
[00:57:31] Annie Chechitelli: That is my hope.
Absolutely. I was saying like, what do I want a year from now? So when you and I are talking a year from now, I'm sure We'll, I would like the tension and the discussions around misconduct to go away. I mean, go away. I understand that. Heat down a little bit to focus on the right things. You know, I say that, but I also know that I'm sure I haven't thought of everything with turning clarity and there'll be some company out there or some industrious student who tries to find a way.
So, I mean, there'll always be something just by the nature of it, but I'm really hoping just for the, a large majority of the normal curve that we can kind of put this aside.
[00:58:15] Alex Sarlin: Yeah, I share that. I mean, it's important stuff and as you said, teachers, you can get the hand wavy stuff. We all should embrace AI and use it and it's important for the future workforce.
And then educators left having to say, what's my actual policy on this? What am I writing on my syllabus? What am I telling students? Like, this can't be hand wavy for me. I need to actually put this in practice. And I think you're really addressing that very directly. So related to that, I think this would be my last question, even though it's a kind of a niche one, but I think it's something I know you'll have a good answer to it, which is that teachers themselves are on this huge spectrum about whether they are using ai, whether they believe in it and love it or hate it and fear it, or whether they think it's gonna take their jobs or whether they think it's gonna ruin their students' education.
It's like you have zero to a hundred with all the educators out there. When you introduce a tool like Clarity and you give the educator the ability to sort of set the rules of the road, does that go along with like. Teacher professional development or recommendations, or are you sort of helping teachers understand the whole spectrum, especially if they're new to this space, to make sure they don't just go in and say, oh, AI sounds like something I wouldn't want off, off, off.
They can't use it for anything. And sort of not even going through that idea of maze of whether to use it or not. I'm curious how you address that problem or how you're thinking about it.
[00:59:31] Annie Chechitelli: Yeah, I think, I'm not gonna say that we're gonna address it perfectly at the beginning at all, right. I think that we don't even know a lot of this, but we do always launch with resources.
We've got a team that we're former teachers, so we always launch things with those resources, and especially with this proc, we started with the AI indicator product. We do a lot more in-app guidance. Gone are the years of like, lemme go read a man. Like no one's reading manual, no one is reading a newsletter.
I mean, maybe, I dunno. I dunno who those people are. And so that's one of the things that's actually a different conversation someday we should have, how we reach educators is really hard. Even for good stuff. Like all of it, like, it's just so hard for us to communicate with educators with stuff that we think would help them.
And I would love thoughts and advice and a conversation on that topic because the old way of doing it, of like, oh, let's just go email the admin and the admin will tell the teachers like, it doesn't work. So we're really being thoughtful about more in-app guidance and discovery and figuring out through some like different testing of how we're gonna do that.
I would expect that to be like a, we're gonna do some basics this year, but I would expect that to be a focus for us in 2026.
[01:00:49] Alex Sarlin: Yeah, no, it makes a lot of sense. I mean, they talk in product design as you know about like good defaults, right? What is the default behavior of what is on and off? And then when they go to.
Examine it and toggle different parts. How do you sort of help them understand the ramifications of that while they're in the app at the moment of choice, rather than it being like a manual or a video or a newsletter or a YouTube channel that they have to sort of go through all this themselves? It makes a lot of sense.
I mean, I have a couple ideas about how to Okay. But we, we can talk about that offline. But I agree with you. I mean, it's really hard. It's different than some other professions in that there aren't sort of mandatory memos to everybody that they Absolutely. I'm sure there are some where there's a lot of information swirling around.
Much of it is very optional, much of it is very peripheral, and teachers are already overworked just writing. They, the 54 hours a week median amount of time that teachers spend on their work, that's what they report on. So like they're already so overburdened. The idea of then adding additional PD or additional suggestions or recommendations or newsletters or videos is just, it seems like out of control.
So it's an interesting question to tackle. I'm really looking forward to turning in clarity and. Whatever, how the entire field evolves to getting to that point where suddenly the graphing calculator moment where people say, you know what? If you're in AB calculus, you can use a graphing calculator and it's allowed on the test.
And those are the rules. You can't use it in fifth grade. And now we all know, and it's accepted and it's across the whole school. It's across the whole college board. It's across the whole AP classes. I feel like that moment, if we can do it intelligently, it's gonna be really clarifying for everybody.
[01:02:24] Annie Chechitelli: I agree.
[01:02:25] Alex Sarlin: Yeah. And avoid, some of this teacher thinks this, but this teacher thinks this. I did the same thing in two classes and I got flagged for plagiarism. Now I'm in huge trouble. Like nobody wants that as a student. Nobody.
[01:02:35] Annie Chechitelli: Not even teachers. No one wants that. We just wanna move past, how do we. How we're developing a generation of people who continue to learn and progress and can solve problems because we got enough of them.
[01:02:47] Alex Sarlin: Exactly. We do have enough problems. I really admire the way you're thinking about it at Turnitin, you're always very careful, like you mentioned the grade scope acquisition. You very careful to not just be the police side of the fence. You have integrity checkers, you can flag at risk situations, but you're also creating formative assessment and grading support for teachers, and now this entire sort of AI playground where students can use it and learn how to use it in a really safe environment.
Sounds really exciting to me. Annie Chelli, chief product officer at Turnitin, global leader in academic integrity grading LMS integration, all sorts of things. Thanks so much for being here with us on EdTech Insiders. Thanks, Alex. Hi for our deep dive this week on week in EdTech, we are talking to Sarah Mossoff, the CEO and Co-founder of Winnie.
And Sarah has been on the podcast a couple of times over the last couple of years. She does incredibly interesting work with Winnie, but they just expanded in a really interesting way that I think is gonna be relevant to our audience here. So welcome to the podcast, Winnie. Welcome to the podcast. Sarah,
[01:03:49] Sara Mauskopf: you can call me Winnie.
This is like my favorite thing. Now some people, when they see me, they slip up and call me Winnie, which is how I know the brand is bigger than me.
[01:03:58] Alex Sarlin: Exactly, exactly. So yes, welcome to podcast. Tell us what you are announcing this week.
[01:04:03] Sara Mauskopf: Yes, so Winnie, which for the past nine years has really focused on helping parents find early education.
So daycare, preschool, things like that is expanding to K 12 education. So you can find not just your preschool, but your. Child school on Winnie and compare all the options available to you. And we are rolling this out as we speak. So this is really real time news.
[01:04:29] Alex Sarlin: Yeah. Some people listening to this might think that makes total sense.
I get why. And others might think, what do you mean? You know, searching for your K 12 school. Don't people usually just go to their local public school? Isn't that how it works? Feels like the system has changed pretty dramatically in the last few years. Tell us why you're expanding in this way.
[01:04:47] Sara Mauskopf: Yeah, so I mean, that was our assumption too when we started winning nine years ago.
Like everyone just goes to their local districted. Public school education is a public good, as it should be. And we always felt like, darn, it would be great if childcare was that way too. Like, and there have been moves in that direction actually with more public pre-K. Yeah. My state, California now has public pre-K for all four year olds, which is exciting.
And one thing we noticed is that as we're getting more public options for early education, we're also getting more alternative options, both public and private. For K 12 education. And so, you know, one public option is charter schools and these have really exploded recently. They are public, but they are not, you don't necessarily just go to them.
You have to apply and they may not be, you know, right next to your house. Like, like a public school might be also kind of emergence of other alternative forms of K 12 education. Homeschooling, micro schooling, independent schools, religious schools, all of these have really gained in percentage points. So what we were finding is that it is no longer the case that you just know where you're going once you start kindergarten.
Parents are actually trying to understand what options may be available to them, what can they afford, what funding might they have from their state with things like ESAs and School Choice, and we wanna be there for parents as they're understanding all those options.
[01:06:20] Alex Sarlin: Let's talk about the ESAs 'cause I think that's a really interesting and relevant movement to what you're talking about here, right?
ESAs have grown in, I think they're in at least 12 to 18 states, something like that, maybe even more at this point. But the idea that parents actually have a education savings account and ESA at which gives them some of the money that would have gone to their public schooling to actually be able to spend.
On their child's education in a variety of different ways. It basically gives them a, an account and some choice there, and that changes the game entirely for how do they know where to spend it and where to send their kids.
[01:06:54] Sara Mauskopf: ESAs and school choice, like it is a very controversial topic in some, you know, you ask some people and they'll say, yes, this is wonderful.
More choice, more funding for education. You ask other people and they'll say, this is horrible. This is defunding the public school system, and I think regardless of whether you think it's good or bad or right or wrong, the reality is that in many states it is expanding and really kind of exploding in recent years, and now there's even a federal mandate to say like, we want, you know, more states to roll out school choice.
I think it remains to be seen how many more. Will due to that federal mandate or just due to the politics in that state. But I think it can't be ignored that more and more states are rolling out these policies and then even in places that are very quote blue and liberal places like Washington DC, you may not have school choice per se or ESAs, but I think it's something like 40% of kids in in DC go to public charter schools.
So you have it really a form of school choice, whether you call it that or not. And that is public education, but it, it is different and you need to apply and understand your options. And so I think we're kind of seeing it from all directions, from the red states, from the blue states, and this movement is happening and so, you know, we wanna be there to guide parents through it.
[01:08:21] Alex Sarlin: Yeah. You also mentioned that some new. pre-K and K two programs, there's basically a expansion downward. You mentioned, right, where people are starting to, 'cause you've been in pre-K and daycare for a while, but now there's a little bit of luring of the lines where more schools are often continuous services.
Tell us about that.
[01:08:39] Sara Mauskopf: Yeah, so one thing we also thought was true nine years ago when we started is like there's your preschool and then you stop and you go to your elementary school and like there is a clear division there. But what we're finding is, is more and more states roll out a form of public pre-K.
A lot of times that takes place inside your public, local, elementary school. And so, you know, in California you get your 4-year-old year for free. Now if you go to your local public school, which is wonderful, and I know many people that take advantage of that amazing offering, and hopefully it will even expand to younger grades, but what that means is the decision point is moving up.
So you're not necessarily making that decision point of public versus private versus charter versus whatever. At five years old, you're making it younger and younger. And so again, it's like this blurring of the lines. There's not this clear division of like, Winnie should just work for these ages because you may actually make a decision early on that impacts where your child goes through fifth grade or eighth grade or 12th grade.
And so you, you really do wanna see the full landscape as you're making those decisions.
[01:09:47] Alex Sarlin: So speaking of the full landscape, tell us about some of the factors and filters and information that parents can use when they're in Winnie to make sense of all of the different school options. Uh, you've done this for a while.
I'm sure you know a lot about how parents like to split and think about what they want and how close it is, what they offer, but I'd love to hear you talk about it. I'm sure there's things none of us have thought about.
[01:10:08] Sara Mauskopf: Yeah, so one of the things we kind of dove into with K 12 is like, how do parents make decisions or what factors are important to them?
And we found it actually is very similar to early education in that there is not one score or factor that matters to all parents. And I think some of the tools out there are about scoring and rating and ranking. Schools against each other. And that really doesn't make sense because for some parents, what matters is the community and what is the community like?
Is there parent participation? For other parents, it's the sports teams. As you think about, you know, middle school and high school sports really matters to some families. For other parents, it may be location and
[01:10:51] Alex Sarlin: or special ed or English language learners support. Yeah, special ed
[01:10:54] Sara Mauskopf: is a huge thing that comes up, especially as your children get older and you actually understand there are differences in how children learn and their needs.
And like, you know, there's a wonderful school near me that focuses on children with dyslexia. And like that is only gonna be relevant to very specific families. But for those families, that is probably one of the most relevant things. Yes. And so again, it really matters kind of understanding holistically what the school's about, not just a score or a ranking.
And then it also is a search you may do more than once. Per child or more than once in a child's, you know, schooling career. You may, something may work now, it may not work later. Something may work for your older child and it may not work for your younger child. And so as we kind of talked to parents and learned about this, we realized you really do need a Winnie for this.
And so like, Winnie should be the this,
[01:11:48] Alex Sarlin: right? I mean, you think about the idea of the schools are good in a place and you have places like, I mean, niche in grade schools and that, as you say, like create grades or numbers for schools. But it's like the US news and World report, which, you know, nobody really believes it or likes anymore.
The idea of reducing all of the different characteristics of a place to a single ranking obscures so much about what's actually going on in any given school.
[01:12:11] Sara Mauskopf: And I think, you know, to the people who are worried about ESAs and school choice kind of defunding public education, I think assigning a ranking or rating only kind of further disadvantages, a public school system that has a ton to offer, which may not be captured in a ranking or rating if you're not focused on moving that number, paying for upgraded ranking or rating.
And so, you know, we're really about making sure you see the full landscape.
[01:12:39] Alex Sarlin: Yeah. So one thing I always find interesting about these changes, right? You said the rise of charter schools, the rise of ESAs and school choice in some states is that it creates a scenario in which if parents have choice about where they're sending schools, then arguably schools have agency about recruiting or marketing or showcasing their strengths to parents.
And you can imagine why they would wanna do that for a lot of reasons. Do you see that beginning to happen now or is it already happening for a while? I know private schools have done this for a long time, but do you see charter schools or public schools or kindergarten or pre-K programs starting to be like, we wanna brand ourselves in a certain way so we can attract all the parents of a certain type?
[01:13:19] Sara Mauskopf: Yes, and this was also something that kind of forced us in this direction. We noticed a few of our customers were actually. K 12 charter schools, public charter schools. And we were like, how did they even sign up for Winnie? Why? Why are they paying us? And what we realized is that some of them had a pre-K and we actually also list kindergartens for some reason we didn't go up before this, but we had kindergartens and they do really wanna get in front of kids at a young age and market themselves and market their offering and really share what makes them unique.
And independent schools, you know, what we call private schools, have always been really good at this marketing. What makes them unique? So, yes, and I also think it's the kind of environment has gotten a bit more competitive. Used to be the case. Even, you know, in the childcare market where there were long wait lists for a lot of these businesses, and what we're finding now is that, you know, actually the market is becoming a little more demand constrained.
You can't just be any old business out there. Parents are discerning. They care where they're spending their money, especially when they're paying for it. And the businesses really do have to market to attract the families that are relevant for their program and explain what makes them unique and in some cases, tailor their services to what parents are looking for.
We're seeing more early education preschools roll out things like aftercare because parents. Are not happy if they have to leave work before the workday ends to pick up their preschooler because it's a part-time preschool. And so I think it's good overall when the supply listens to what the demand needs.
It's positive for families as long as there's options that families can afford. I think that's the big thing that we wanna make sure there's enough of, which is that there's enough public options that are competitive and catering to the needs of families and children, but that there are enough to meet all the demand.
[01:15:18] Alex Sarlin: I mean, I know during the pandemic we saw over a million students leave the public education system in the US and go in a variety of different directions and you know, you mentioned micro schools in passing before, and it strikes me as I hear you say, it's like micro schools not only have to market themselves, they basically have to explain what they even are.
I mean, this is the type of schooling people have hardly ever experienced in any way. Very few parents went to a micro school. So micro school probably asked to say, this is what it is, this is how we work. Tell us about how that looks on Winnie and where you expect that to go.
[01:15:49] Sara Mauskopf: Yeah. I mean, I think right now we don't list homeschool options on Winnie because there's not a way to really show that.
But I could imagine over time, maybe at least explaining to people that the differences or the kind of curriculum you might institute in homeschool environment. I think micro schools are. Interesting because that's actually something which does have a physical location. We could list some on Winnie parents nearby could discover them when they're searching.
So those we do aim to include. It's a little challenging because they're not all in the public databases, so we have to. Make sure to grow that data over time, but I think that's a reasonable challenge we're up for. But I think what's interesting with the changing landscape is even pre covid homeschooling was very, very niche.
And it's still pretty niche, but it's growing like it's over 5% now of kids are homeschooled. And I think there's a lot of things that would drive a parent to homeschool, but many may be looking for something in between, like fully homeschooling, it's all on me and sending my child to a public school. And that's where I think micro schools come in, potentially.
Online schools, there's one-on-one schooling, there's alternative schools, there's all these kind of things in that, in between that. I think are now becoming more relevant.
[01:17:10] Alex Sarlin: Yeah, and I mean, nobody knows exactly how it's gonna evolve, but I can imagine a world in which homeschool parents who have been sort of banding together in some ways and creating these micro schools start to say, okay, well I want to continue doing this even if my own child graduates out of it.
Or even if just sort of, how do you think about the future of schooling in that way? You have all of these people who have now sort of deemed themselves and legally been able to deem themselves as the school leaders. But you know, if you're down the block and you wanna know whether that micro school is a place you would actually consider sending your children, there's just no way to find that out.
[01:17:45] Sara Mauskopf: Yeah. I think there's a lot of similarities between micro schools and home daycares, which is something we. We really credit ourselves, not necessarily everyone else crediting us. We credit ourselves with putting home daycares on the map. People did not use home daycares a decade ago when I was looking for childcare for my kids.
They were really, especially in. Really affluent areas like it was not considered something that someone like me would send their child to. That has completely changed, and a lot of parents these days are seeking out home daycares as they should because they're an incredible option. You can get a really small environment.
Sometimes on your block there may be a home daycare.
[01:18:30] Alex Sarlin: Exactly.
[01:18:30] Sara Mauskopf: And when you can discern quality and really understand if this is a good option, it becomes something that you might pay the same as a childcare center, if not more for. So that has really changed. And I think micro schooling could be similar, where like, yes, you need to be able to discern quality, they need to be regulated, you need to be able to find them.
But when all those things are taken care of, it could be a really great option, especially for certain kinds of learners.
[01:18:55] Alex Sarlin: Yeah, it makes sense. It's a really, it's just such an interesting fracturing of the traditional environments like where, you know, new options coming up, sort of grassroots options coming up.
And as you say, that's happened in daycare for a while. As the costs of daycare and childcare has spiraled like crazy, which we've talked about in previous podcasts, it actually makes economic sense for home daycares. They become a really viable option if, as you say, there's some quality control regulation, it's not just fly by night 'cause there could be risks there.
And I think the same is true for micro schools. You can imagine some incredible micro schools and you can imagine some micro schools you would never wanna consider sending your kids to and without the ability to that quality from afar, it's very hard to know if that's even an option.
[01:19:36] Sara Mauskopf: Yeah, when we started Winnie over nine years ago, there was this assumption in the market that like if you're affluent and you have a baby, you would hire a nanny, and if you're affluent and you have a baby, you would not send them to a home daycare, which is really crazy to me because the nanny market is completely unregulated.
You're just hiring someone off the street who calls themselves a nanny, whereas the home daycare market is actually regulated. You're required to be licensed in every state. Plus now with things like Winnie, we actually have quality measures in place, and parents can review them and you can see exactly, you know, learn more about the business.
And I think homeschooling and micro schooling, there's some similarities there. Like homeschooling, you're really on your own to assess quality and it's you being the teacher with potentially no or very little support. And some people may be excellent at it and others not so much. Whereas a micro school could become more like a home daycare where it's regulated and you can assess quality.
I don't think it's there yet, but I could imagine the market evolving that way. And then again, you have these small business owners and I think more small businesses are good for the overall economy and market.
[01:20:46] Alex Sarlin: So last question, a little bit of a business logistics question, but I'm really curious about this.
Winnie is a marketplace, right? It's a two-sided marketplace, and it's a marketplace where one side is parent consumers. And I think, you know, a lot of our listeners are founders and entrepreneurs, and there's always this sort of debate in EdTech about direct to consumer models versus selling to schools.
And you sort of have a little taste of both, right? Yeah. Which is kind of interesting. So I'm curious if you could just talk for a bit about what you've learned in the almost decade about parents as customers and how to control acquisition costs and things like that there versus schools as customers and how you get on their radar and sort of break through the bureaucracy.
[01:21:26] Sara Mauskopf: Yeah, I mean for us, when we were first starting out and trying to monetize Winnie, we kind of tested both sides. Like do we charge parents a subscription for access to Winnie, or do we charge the businesses to kind of have enhanced tooling, better marketing, and we could not make charging parents work, like parents are really stretched thin when it.
Comes to all the things they already have to pay for. Including childcare. Yeah. Charging them more. On top of that, we didn't see a future that way where we could get to scale. However, charging businesses when we can make their business more efficient and grow revenue for their business made a ton of sense.
Like if we can fill model.
[01:22:05] Alex Sarlin: Yeah.
[01:22:05] Sara Mauskopf: Yeah. Like an open space in a daycare is, that's just profit. If you're leaving that open and we can fill it for you, you're just, you've already. Paid all your fixed costs. And so it was really natural to sell to the daycares and preschools. And we think similar for schools, like they're just even beyond just the marketing component, there are so many ways to make schools more efficient with tooling and ai.
And so like we think those are our customers, not the parents, but obviously we need to attract parents, right? Really challenging because parents are the consumer segment. Every single product is going after whether or not it targets parents or is just a car manufacturer. So it feels like a really easy segment and investors are always like, oh, couldn't you just post it in a mother's group?
And like, no, because there are millions of mothers groups, word of mouth with parents is actually really, really hard. So do not underestimate that going direct to the consumer parent audience will be incredibly challenging. There are certainly brands that have done it, but it's very, very hard.
[01:23:16] Alex Sarlin: So how do you get parents to know about Winnie?
[01:23:19] Sara Mauskopf: Yeah. I mean, that's why we're still here. Still kicking. Yeah, still working on this almost a decade later because it's a hard, and that's why I do podcast,
[01:23:28] Alex Sarlin: try to reach
[01:23:30] Sara Mauskopf: more of our audience. We have a huge advantage in that childcare and education is something parents are seeking out information on, so we don't have to be like, here, look at our cool stroller.
Like you should definitely buy this. They're actually typing into Google, daycares near me, schools near me, all these niche terms that we now rank for, which is great. And then over time, we've just built our brand and our audience, which is wonderful, but it was not. Easy and it takes so, so long to build a brand.
And so like I don't recommend that paying to anyone. So basically like build the brand. Yes. But there's other ways to monetize other kinds of payers. Sometimes going that route where it's someone is already used to paying for and you're just building a better thing. Sometimes the easier path, at least for us it was.
[01:24:27] Alex Sarlin: That's good advice. Yeah, it's really interesting to hear you think that through. It makes a lot of sense. Well, unfortunately we're at time, but this is so much fun. So where can people find Winnie in the app store or online, and how can they learn more?
[01:24:39] Sara Mauskopf: Yeah, so winnie.com, W-I-N-N-I e.com, or in the app store search for Winnie.
And yeah, we're on all the social media channels. You can't miss us.
[01:24:52] Alex Sarlin: I've used Winnie to look for daycares, but now I'm going to start expanding and looking for kindergartens and starting to understand that better, and I'm excited about that. I think it's a great interface. It's really easy to use. It's a great product.
Thank you so much. Thank you, Sarah. Thank you. CEO and co-founder of Winnie expanding into K 12. Thanks for being here with us on EdTech Insiders.
[01:25:11] Sara Mauskopf: Thank you so much.
[01:25:13] Alex Sarlin: Thanks for listening to this episode of EdTech Insiders. 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 Ed Tech Insiders Newsletter on substack.
