Priten: Welcome to Margin of Thought, where we make space for the questions that matter.
I'm your host, Priten, and together we'll explore questions that help us preserve what matters while navigating what's coming.
We spend a lot of time asking how teachers should respond to ai, but what about the people actually building the tools teachers are being asked to use?
Today's guest is Yani Chen, a master's student at the Harvard Graduate School of Education, studying learning, design, innovation and technology.
Yani is also a product developer at Deep Brain Academy, where she's designing an AI math tutor with something rare in ed Tech, A genuine commitment to scaffolding, cultural inclusivity, and keeping teachers in the loop.
We're gonna talk about what it actually takes to build AI tools that support learning rather than shortcut it.
Why Gen Alpha's fluency with AI makes the educator's role more important than ever and what a student and product developer sees when she looks at where education technology is headed.
Let's begin.
Yani: So, Hi, I am Yani Chen.
I'm currently a master student at Harvard Graduate School of Education, and I'm studying learning design, innovation and technology.
Priten.: Do you have some experience in the field already?
Um, so can you tell us a little bit about what that experience has been?
Yani: Yes, I do.
So in the past I've worked as a community manager for a EdTech company called Lobster, and currently I work for a EdTech startup called Deep Brain Academy to work on their AI educational product development.
Priten Shah: So on this show, we've heard from a lot of different folks, um, but we have yet to speak to a company that's developing, um, science
ed tech tools., I'm very curious to hear your perspective on what responsible, , development of these tools looks like and the utility of them.
If you wanna start generally with how you approach, . When is EdTech, um, useful in the classroom versus when do you think that it actually distracts, , from the class?
Yani: Yeah, so for the product I'm currently developing, I think one of the major challenges that we encounter into is how can we take teacher into
the position aligning with ai, because you know, like for the Gen Alphas, they're already really fluent AI users and they're the most fluent users.
Of any generation.
So it's really important to get to know that how teachers could be in the position that helps those gen alphas to develop their process.
So one of the things I do consider is about the ethics of ai.
Because AI are trained based on the broad database that includes a lot of biases.
So we need to take consider as teachers to let the students to know that, how can we navigate this kind of situations.
Priten Shah: So I'm curious in terms of, , how you navigate both, , trying to get uptake for the product and get folks to use it, um, and get excited about it, while also talking about some of the potential dangers, like that balance, um, , is tricky.
Yani: Yeah, so some of the things I would let my students to use is such as the one that I'm creating now is a math tutor, and students could use this math tutor AI to work on their math problems at home or at anywhere where like teachers couldn't follow up.
However, there's also the problem with AI is sometimes they do have some logistic problems and also like ethical problems by giving math examples.
Like for example, in the classroom, they'll give like math examples that are not culturally inclusive and we don't want students to experience that kind of example because that will frame their work to a biased way.
So when we're training the ai, we about putting as inclusive materials as possible and all the kinds of like pedagogies and curriculum design frameworks to let the students to understand the knowledge, but also understand the ethics at the same time.
Priten Shah: Is a lot of the work that you do predominantly, um, building tools for use outside the classroom, or do you also, , have projects that are for,, direct use in the classroom?
Yani: So the product I'm creating on have two different version.
One is outside of classroom, but also for inside of classroom.
We're thinking more about the safety issues.
So like if students do input any like dangerous messages or inappropriate messages, they'll alert.
The teacher and the teacher will come to the students to help to navigate out the situation.
So for the in class, we want to take teachers like as into the process as possible.
So when AI is answering students' question, they'll only give like step-by-step guide, but not giving the direct answer.
And this case, if the student is still struggling, the teacher will come and help out the student.
Priten Shah: Have you already heard from teachers about whether , they're excited about something like this?
Is it daunting?
, Because we speak from the other side of, you know, going into schools and helping them think about which tools they might use or, , whether or not they even use a tool.
, I wanna know, , from your experience, , are teachers like looking for something like this and are they, um, excited about it?
Um, and What about it, uh, excites them?
Yani: I feel like teachers are mostly holding skepticism towards AI because they're not trusting AI that.
as much.
However, as soon as the teachers uses AI in their classroom, they're kind of more comfortable with navigating this kind of tools.
I feel like it's just the same thing as we use like PowerPoint slides, we use videos, those multimedia, um, things that we used in the past.
It's same with AI right now.
It's just a new kind of tool that helps teacher to teach students better.
Priten Shah: Do you see that shift happening with teachers once they start, once they get some exposure to the tool?
Yani: Yeah, I do see some shift, but like most teachers still hold skepticism, but like I do get valuable feedback from the teachers about like, how can we develop our product even more
Priten Shah: I'd love to go into the details of what that process looks like.
Because I think one of the things that, We, we hope that more ed tech companies do is stay in conversation with teachers build tools that the teachers want and not, you know, make the teachers want the tools that the ed tech company is building.
So I love that you all are getting that feedback and iterating based on it.
Is there a formal process for that?
Does that end up being a, a lot of,,, ad hoc conversations?
, Just want to know a little bit about how you've managed to, uh, make that happen.
. Yani: Yeah, for now we're doing mostly conversation.
We haven't had really a formal evaluation because there isn't a lot of teacher that are willing to adapt these kind of tools in their classroom as most teacher are more like fixed to the traditional classroom style.
So uh, one of, like, the most common feedback that we give is that students get addicted to AI tools when they're using classrooms because
ai, they could like give out information so fast, which teachers sometimes have to like apply to each student and make it personalized.
But AI could do that in seconds.
So student will ask all kinds of questions to ai.
So that's also like the one thing that we develop later is to add safety checks, which is like red teaming, just to let students to not get too much with ai.
So when student ask any irrelevant questions, we will pause them for a second and let the teachers to come back and say , okay, you should do your math work right now.
But not like chatting with AI about, some random sports game or like any like concerts.
Priten Shah: very cool.
So, So the goal is you to like, uh, initiate human intervention actually at those points and kind of get the teacher to come in and, and, uh, get them to get back on task.
That's super interesting.
so I know you're, You're talking about how skeptical teachers are right now, , and some of the, the friction that you're facing.
What keeps you motivated to do the project?
Yani: I'm really excited how AI tools could actually reduce a lot of time for the teachers.
'cause like teachers have so many workload right now, they're overstressing themselves and just by applying AI tools, they can reduce like a lot of the time to.
Do like other kinds of work because like in the past when teacher goes to tutor student, they have to take the full hour just to talk with the students, step by step, guide the students to think about the math problem.
But now student could do that at home and in class, teacher could more focus on the content to get student engaged and to get them interested into the knowledge that we're talking about.
Priten Shah: I know you said you're also a student yourself.
Um, and so I'd love to speak from that a little bit.
What motivated you to go back to school?
Um, what role are you hoping that plays in your future career?
And then I want to hear a little bit about your experience as a student.
But let's start with the why.
Yani: Yeah, so I feel like because AI is developing so fast, we could never study more enough of ai.
So going back to school really helped me to learn about the most emerging tech.
that we have in educational space and just to spark new ideas about projects, about how to let teachers to teach better and also how to make the student more engaged in classrooms.
Priten Shah: Tell us a little bit about the program.
You're not just doing a Master's in education, right?
It's, it's a specialization, , in a very relevant field, though, as you wanna explain the field for folks.
Yani: Yeah, so the FOLK program is called Learning Design Innovation Technology.
So most of my coursework are focusing on like, developing AI agents on curriculum design.
And also like using multimedia and one of the course I'm really in excited about right now is using laser cuttings 3D printing to do really like sense making objects for students to learn STEM knowledge.
I feel like that's a big part of learning is that you do to build something and you build, you learn from building.
Priten Shah: Is this to get teachers to build, tools for the classroom or to get the students themselves to build, um, with the printer?
Yani: It is for the students to build themselves with the printers.
It's really important to teach students those techniques as they can, like spread out their imaginations of the objects they want to build, but also we have to be cautious about, we're not doing like mass productions, , objects.
So like there is one of the cases that I read about is about creating.
Acrylic key chains that students are really obsessed about and it becomes mass production, and we don't wanna design a curriculum like that.
Priten Shah: So how do you teach that intentionality, to students to get them to, um, you know, really consider what they're building and the implications of what they build?
Yani: I feel like it's more focusing on what is the intention of designing the course, because we want students to learn the technology and to learn the content behind the building, so it's more of designing.
Not too hard, but also not too easy, adjust right curriculum.
And it's really hard right now 'cause I feel like I'm still in the process of um, navigating this process.
I still haven't figured out what is the right way to design for it.
But through the feedback from student, I'm really excited about what is going on.
Priten Shah: Very cool.
What role has AI played um, in your own education?
Yani: I feel like AI really helped me a lot because during courses as a student, I often take pages of notes, but when I look back and want to revise of the all the content, I couldn't really kept the focus point.
So like I would just input it into ai, let it summarize it for me, and I'm ready for the next class.
Priten Shah: So obviously when folks are talking about the studying process, one of the concerns is that by having AI do that synthesizing or summarizing, um, you might be missing out on some of the learning.
It sounds like that wasn't the case with you.
You, You actually benefited from it.
What do you think led to it being helpful for you rather than just like a shortcut that didn't end up showing benefits.
Yani: I feel like for me it's really important that you have to learn for yourself 'cause you are not learning for ai, you're not learning for anyone.
knowledge that you absorb is totally yours.
So during class I will pay as much attention as possible, and then when I go back, I want to revise the content.
I'll use AI to summarize the bullet points for me so I will get a clear picture of what I did during the classroom and what is going on with the examples and the scenarios that we are discussing about.
Priten Shah: now That's a very independent use of the AI technology.
How do your professors feel about it?
Are they talking about Um, using it in the classroom, are they, do they have restrictions, um, whether or not you're allowed to use AI in certain ways?
Especially because you're in a learning technology track.
I'm curious how, your professors are approaching it.
Yani: So most of my professor are approaching in a optimistic way.
They allow us to use AI to brainstorm, allow us to summarize and allow us to just to get a clear vision about all the stuffs.
However, there are like strict restrictions on using ai, such as we're not allowed to use it for our assignments and we're not allowed to use it for any of the final product that we created on.
But still, I feel like AI is really useful in the brainstorming sessions 'cause you can't think of so many ideas from yourself, but AI could help you and guide you to the things that you want to do.
Priten Shah: When you say that they're not allowing it in the, like the final product or the final assignment, is the idea that you can like use it as a
thought partner and come up with the idea, but the full execution has to be yours or, are there other places during the process that you're allowed to use it?,
Yani: So we are allowed to use it as a thought partner, but not about like specifically writing the assignment.
But still there are a lot of product that involves us building AI agents and also using AI chat bots.
So that part that would be totally okay.
But for like writing your own thought about your reflection, about discussion of the content is not allowed.
Priten Shah: Has there been a moment where, you were working on a final project and AI made something possible, um, that you would've not otherwise been able to do without ai?
Yani: Yeah, so in my J term class, which is January term, I've been in the course about like building, uh, agents, and it's my first time using Google AI Studio.
I'm really fascinated about like how much function it could do.
It literally creates a whole website with all the functions that I wanted, and I could add individual prompts to all the things that I wanted.
So.
It is really easy to vibe code in these days 'cause AI is just so powerful and it could help you to create the whole product by yourself.
But we also have to be cautious about what information do we want to input into ai and what kind of outcome do we wanna get from it.
Priten Shah: The vibe coding part is what I'm, I'm curious about, um,, do you have a programming background yourself or, um, are you building things for the first time using tech, through, um, AI tools?
Yani: No, I do not have a programming background and it's actually the first time I vibe coded because, and before, I'm mostly using AI for.
Summarization.
I never thought that I could use AI for vibe coding, and it's really helpful for me to create something that I couldn't do before.
And also for the aesthetic part, I'm not a UI but the AI did design it pretty good.
Priten Shah: It's an exciting use case where, when, you know a lot of folks in education, kind of, especially if you're in the classroom and
you're teaching, , or you're working directly with students, you kind of have an idea of, oh, like this, this one tool would be very useful.
Um, , And oftentimes there's a little bit of a disconnect with, um, you know, , what the teachers want and what the ed tech companies are building.
So I'm excited to see more folks who kind of can bridge that gap, , and also build things on their own.
. Obviously vibe coding might not fully be there yet in terms of security.
And so we want to keep, we want to keep seeing, , how that develops.
But it sounds like, , there's some really creative use cases that are coming out of the potential, , to build your own tools.
When you consider, the path you want to take post-graduation, what excites you about like next steps in your career?
Yani: So I'm really excited about becoming like a AI educational product manager because I feel like there's a lot of potentials of AI products, especially in the
educational fields that still haven't been discovered yet and is excited to become a part of it and to become the innovation part of helping students to learn better.
Priten Shah: You know, when we think about the use of AI in education, there's a lot of excitement, that I share about some of the potential use cases.
, But some of the concerns that we often hear are, we don't yet have research that it actually changes how students are learning, right?
So maybe it's a different way for them to learn, but it might lead to worse learning or the same learning, and AI comes at a cost.
there's a financial cost, there's the training cost and time cost for teachers.
Um, , But of course there's also like ethical costs to the, you know, environmental concerns, you talked about, uh, data concerns, So I struggle with figuring out when that benefit to the learning process, um, will be demonstrated, um, but also worth it.
Have you thought about, um, , how you might approach that or just like started thinking about it?
I'm not, obviously I'm expecting you to have solved it, but I was curious where you're, at
Yani: Yes, I do have thinking about it because in the states we actually lack of guidelines of using AI and also.
The things that you mentioned about data, security, about safety, about ethical issues, we lack those kind of guidelines.
So I think it's really important for us as educators to set our own individual guidelines at this point.
But since AI is evolving so quickly, I feel like in the future there will be some kind of, like, systems that come out with the guidelines that allow us to rethink about how to apply AI in classrooms.
Priten Shah: Do you think we'll get, federal or state policy, or do you think that it'll be like, educators in schools and, you know, educational organizations that will help guide this?
Yani: So currently, in companies and in schools, we have like individual guidelines.
However, we do want the federal and state or school district level to have like a reactive rules about guiding how educators should use AI tools.
But like currently, since we're in the early stage of implementing AI into education, I felt like each educator should have their own guidelines about what to use and what not to use.
And there should be clear lines of data security and also safety and also ethical issues about using ai.
Priten Shah: For our educator audience, what are some things that you would recommend from, um, you know, someone who helps build these tools
and thinks about it from the inside that educators should consider when they're deciding whether or not to adopt an AI tool or any ed tech tool
maybe.
Yani: I feel like one of the thing is that you have to think of whether it aligns with your own ethical guidelines.
'cause AI has been trained from a large data set.
It has a lot of different kind of perspectives.
So you have to be cautious about how to navigate those things.
And if you aren't, like, if you aren't ready to experience all kind of perspectives from AI that has been trained from the dataset, I feel like it's not yet the time for you to imply ai.
Priten Shah: and What can they do to develop that?
Yani: I feel like the only thing is by using it to be honest, because the more you use a tool, the more you're familiar with the reaction, the function of the tool.
And by using the tool, I'm not meaning that you're gonna imply in classroom, but just for your daily personal use, and you'll find the goods, the pros and cons of using ai.
Priten Shah: I think that that's one thing that we've also found very effective is, for teachers who are afraid or not yet sure what role AI will play, in the classroom.
We just wanna learn more.
The very low stakes is use it on your own, your own time.
Don't even think about it, like in terms of your students and how you're gonna use it in a classroom or assign it to them.
Um, but, you know, go find a recipe, right?
Like do something very, very, um, you know, personal low stakes that gives you an opportunities to start exploring the technology.
, you know, we, , We've talked a lot about the chatbots and tutoring.
, But you've worked on other aspects of like ed tech and exciting frontiers in education.
Um, You mentioned 3D printing as one of them.
, everybody's talking about AI and, and text-based chatbots, , but there's other cool things happening in the ed tech space that are also innovative.
Um,, What other projects have you heard of or worked on?
Yani: I feel like laser cutting and 3D printing is one of the things I constantly working on because it is really easy to learn how to use them But there are also like cons about the cost of using laser cutting, three E machines.
But you can also like similar like ai, you can use it to create anything that you want.
So I feel like it is a great learning source for student, but also for educators to develop more creative learning plans.
Priten Shah: Is there anything else besides those that is exciting to you?
Yani: Yeah, AR really excites me and VR actually, because like the past company that I work on, they have like a VR simulation for nursing students to, um, manage with patients and also to be familiar with the daily process.
I feel like it's just a great way to experience this kind of things because, you know, like labor costs and also kind of like safety issues.
It's hard for these students who need to be come like practitioners in fields to practice with real humans.
So in this case, AR and VR is a great way to learn from it.
And I know that in a lot of the medical schools, they use mannequins mimic and do like patient playing, however.
The problem with that is the students don't have the conversation, the communication with the patients, which they could not learn to, how to deal with real time interaction or problems that might encounter in real life.
So I feel like it's just a great way to let students to learn in a safe environment and with the fully guided by the instructor.
Priten Shah: Yeah, so like a training grounds, before they're actually interacting with real humans, where the stakes are much higher.
Are there use cases in K to 12 education that you've, , thought about with ar vr?
Yani: Yeah, I think keep working on like a personal open source project, which is to create free AR models for STEM educators to use to teach students.
'cause you know.
Some STEM equipment are expensive and AR could actually be only used on phones or any technology base, so it is easier for students to learn those STEM concept, but also, yeah.
Priten Shah: Can you gimme an example?
I haven't taken a science class in forever, but, I'm thinking, oh, when would it have been, possible to use ar?
Yani: So I've created AR models of Newton's Cradle it is really interesting how student could click like 1, 2, 3, 4 buttons on the AR model, and there will be the numbers of balls dropping down to show the physics process of it.
And since for most of the schools, there might only be a few Newton's cradle, so they have to be in groups to play with it.
But with ar, each students, if they have a technology device they could play with on their own and they could see and maybe just like visualize what is happening and investigate like their own thoughts.
Priten Shah: I haven't explored what the limitations and capabilities of um, AR are, You know, when I've seen conversations in generative ai, one of the concerns is that, the world models are not great.
And so some of the physics simulations and video output, for example, is, is flawed.
It doesn't get shadows correctly, it doesn't get movement correctly, the speed of some things falling, it's not fully accurate, which, you know, limits, , generative AI video production.
, Are there easier ways to control for that in ar?
Is the technology behind it different?
Yani: So when you're creating an AR model, it depends on the type of tools that you're used to create.
For example, I use Blender where they got really good, embedded physics system that you could use to mimic real physics situations.
'cause I, you know, a lot of games use those kind of models.
So I feel like using Blender is a great choice for creating AR models.
However, . I do understand that we cannot fully copy what the real life is happening, but I feel like it is opportunities for the schools without that much of a resource to use this kind of technologies to educate students.
Priten Shah: is there, is there anything else about your, , schoolwork or work that you, um, would love to share with teachers or you think that they should, hear about so that they can know what's, what's coming, what's exciting?
Any concerns?
Um, any of that?
Yani: I
feel like scaffolding is a big part because AI hasn't really learned what scaffolding is.
They only give direct answers to students and also to teachers or to anyone because that is what it is engineered for.
So as educators, we want to scaffold students, so we don't want the direct answer to come out, and that is the part where I feel like AI in the classroom should be paying extra attention, not because we still wanna take the teacher in the space.
Priten Shah: Are you saying that they should, look for tools that do scaffolding or the teacher has to do a lot of the scaffolding alongside the ai?
Yani: I feel like both.
Yeah.
Like for example for the tools, I feel like teachers should choose the tools that do not give direct answer, but give step by step instructions
and like lead students to think and to discuss about the final answer as well as if there aren't that kind of tools, teachers should.
In this cases to scaffold students, to drive them thinking, because without the thinking process, the students aren't really learning about what they're learning.
Priten Shah: In those use cases, , what would you say to a teacher who says, well.
I'm gonna have to do that work.
Why should I even bother um, using ai?
I'll just do it myself.
Yani: I feel like that's the cons of using AI because AI is still developing and for those kind of AI tools haven't been really in public yet.
So right now when are teaching things like math, AI will still give like direct answers and that's the part where teacher has to come in.
Priten Shah: Yeah.
So it sounds like you're excited about where the technology will be in a few years or maybe even months at the pace things are moving.
But maybe we're still not exactly there where it'll make teachers' lives fully easier or teachers can fully rely on it.
Does that sound, is that an accurate characterization of what you're saying?
Yani: Yes, I do think that because I think nothing can replace what the teachers do in a classroom.
It's about the environment, about the engagement, about the learning process.
And what AI do is it can only output information for the student, but it cannot copy what the teachers are doing for the students to guide their thinking process.
Priten Shah: Yeah, so I'm thinking back to the early example you gave of the math tutor um, where it would be maybe helpful for the students to get some feedback and practice like as a rote, practice tool outside of the classroom.
, But maybe they,, they still have to engage in the side of the classroom with direct teacher instruction.
Because so much of the conversation right now is about ai, I think thinking about all the different ways that, we can actually, not just go to technology for answers and shortcuts, but to actually use the technology, to explore like creativity.
Um, and, you know, kind of go down rabbit holes of learning, um, is super interesting.
And so I think that the 3D printing example is definitely, very cool.
Which classes do the 3D printing, use cases normally come up and is it normally in like a science classroom or is it like art classrooms and um, that kind
Yani: I think it's normally like digital fabrication classes.
If it's in K to 12 settings, I think it will mostly appear in like design classes, art classes, but there could still be some projects in like STEM classes that allow students to work on their own sense making process.
And for humanities, I feel like student could also create a lot of stain by using those kind of tools.
Priten Shah: uh, My last question for everybody is, when you think about the next five years of education are you mostly scared or mostly excited?
And why?
Yani: I think I'm really excited about what's coming up because like AI is evolving.
Not in like years, not in months.
I feel like it's in weeks.
Each week AI will be different.
So for the next five years, I can't imagine how AI would be, and I could imagine is that ai, the introduce of AI would let a lot of people be really convenient of using the tools and yeah.
Priten Shah: I think that that there is some excitement to seeing all these different use cases, especially in stem, I think, about where we can, get our students to be even more hands-on with the real world.
I, I really appreciate your time today.
Um, , It's nice to hear, , from the side of the engineering of stem, um, ed tech tools, particularly because I think there's, there is something
very different about, , using like ed tech in technology or science or math classrooms, , than it is in like English or , history classrooms.
Obviously the,, the skills are different, the context is different.
, And so I think even though there's a lot of fear around it in those contexts, I think it's nice to hear some excitement about it, , in the STEM world.
Priten: Thank you to Yani for joining us and offering a view of AI and education that we don't hear often from someone who is both a learner and a builder at the same time.
Yani reminded us that the design choices behind these tools are never neutral, and that getting them right means centering, equity, scaffolding, and teacher presence from the very beginning.
Her work is a reminder that the future of EdTech doesn't have to be something that happens to educators.
It could be something that we build with them in mind.
Keep listening as we continue exploring the ethics of education technology.
And don't forget to pre-order my upcoming book, ethical Ed [email protected].
Thanks for listening to Margin of Thought.
If this episode gave you something to think about, subscribe, rate, and review us.
Also, share it with someone who might be asking similar questions.
You can find the show notes, transcripts, and my newsletter at priten.org.
Until next time, keep making space for the questions that matter.