Irreplaceable: How AI Changes Teaching & Learning with QuestionWell's Maya Bialik | GSVtv Podcast

Summary

---

Podcast Summary:

ASU+GSV Summit | Irreplaceable: How AI Changes Teaching & Learning with QuestionWell's Maya Bialik
Date: July 20, 2026
Host: ASU+GSV
Guest: Maya Bialik (Teacher, Researcher, CEO / Co-Founder of QuestionWell, Author)
Also Featuring: Meredith Rosenberg (New Advisory)

Episode Overview

This episode, recorded live at the ASU+GSV Summit 2026, delves into the evolving role of artificial intelligence in education. Featuring Maya Bialik—teacher, cognitive science expert, CEO of QuestionWell, and author of How AI Changes Everything and Nothing in Teaching and Learning—the conversation focuses on how AI is transforming classroom pedagogy, teacher agency, assessment, and human connection in learning environments. The discussion unpacks the paradoxical effects of AI—what remains irreplaceable about good teaching and learning, and what AI can (and cannot) change.

---

Key Discussion Points & Insights

1. Genesis & Evolution of QuestionWell (00:30–03:31)

• Maya shares the backstory of QuestionWell, which began in 2023 when she was a middle school science teacher seeking high-quality, learning-aligned assessment questions.
• Initially developed by her and her husband, QuestionWell uses AI to create and assess educational materials, focusing on teacher agency and quality.
  • “The AI writes the questions and it helps to reason about the quality of the questions and what their coverage is and what the bloom's level is. And then the teacher can export them to their favorite platforms.” (B, 01:21)
• Maya emphasizes the shift from teachers as creators to curators with generative AI, but underscores the risk of undermining teacher expertise if AI is blindly followed.
• The tool integrates learning science principles and aims to bridge the research-practice gap in real-time classroom decision-making.

2. What Stays the Same, What Changes? (03:31–04:52)

• Despite technological advances, fundamental cognitive science principles in learning remain constant.
  • “Everything has changed. Nothing has changed. So the core cognitive science of it has not changed. There's still cognitive load, there's still extraneous cognitive load and intrinsic cognitive load.” (B, 03:48)
• AI can offload extraneous cognitive demands and help scaffold or challenge students, but deep teaching practices are still “irreplaceable.”

3. Navigating AI Polarization in Education (04:52–06:46)

• The panel discusses binary attitudes towards AI—excessive hype vs. existential fear—and how the conversation in Maya’s book seeks a more nuanced middle ground.
• Maya introduces four overlapping purposes of education:
  1. Academic (disciplinary tradition)
  2. Humanist (human-to-human connection)
  3. Social (civic responsibility)
  4. Pragmatic (workforce preparation)
• Each lens reveals unique risks and opportunities in adopting AI.
  • “All four lead to completely different views of AI...Every single one shows different risks and opportunities in AI.” (B, 05:39)
• This multifaceted view helps calm polarized debates by validating diverse concerns and hopes.

4. Patterns in AI Skepticism and Adoption (06:47–09:15)

• The hosts inquire about which groups are most wary of AI.
• Maya notes skepticism is widespread and complex, not easily mapped to age, geography, or learner type, but observes a notable U.S. hesitancy compared to global contexts.
  • “In the US we tend to be a little bit more negative compared to the rest of the world.” (B, 07:24)
• Social and humanist camps (focused on connection and community) often voice concern about AI’s potentially isolating effects.
  • “Of course, you could see AI as dismantling that. At the same time, I'm very excited about uses of AI that enhance.” (B, 08:17)
• Pragmatic concerns—on AI’s impact on jobs and credentials—are especially amplified by U.S. media focus.

5. Educator Reactions & Shifting Mindsets (09:15–10:25)

• Maya describes feedback on her book as “informed hopefulness.”
  • “They don't just feel blindly more hopeful. They feel like informed more hopefully.” (B, 09:44)
• The approach encourages educators to hold complex, sometimes contradictory truths, creating a more nuanced, less cynical outlook.

6. Improv, Collaboration, and ‘Yes, And’ Thinking in AI (10:25–12:31)

• Maya’s background in improvisation parallels her optimistic take on AI’s adaptability and collaborative potential.
• Improv’s “Yes, and” ethos can inform classroom use of AI to mediate conflict, bridge disagreements, and synthesize perspectives.
  • “A core principle of improv is Yes, And of course… one thing I'm personally very excited about is AI being able to bridge difficult collaborations and create mediating artifacts that help...people who don't agree to work together collaborate and come to more of an agreement.” (B, 11:07)
• Anecdote: Maya describes a Slack debate where an AI assistant offered context supporting both sides, helping her and her co-founder/husband synthesize ideas and move forward.

---

Memorable Quotes

• On Responsible AI Integration:
  “We're actually building in the affordances for teachers to make informed instructional decisions in that moment of instructional decision making.”
  — Maya Bialik (B), 02:18

• On Enduring Foundations:
  “What remains irreplaceable? Well, among other things, one thing that remains irreplaceable is the cognitive science underneath teaching and learning.”
  — Maya Bialik (B), 04:25

• On AI Polarization:
  “It is both good and bad. And it's up to us and what we do with it to make it more good than bad.”
  — Maya Bialik (B), 06:47

• On Human-Centric Concerns:
  “Of course, you could see AI as dismantling that. At the same time, I'm very excited about uses of AI that enhance.”
  — Maya Bialik (B), 08:17

• On ‘Informed Hopefulness’:
  “That's so weird. I feel more informed, but also more hopeful.”
  — Maya Bialik (B), 10:13

• On Collaboration and AI:
  “AI being able to bridge difficult collaborations and create mediating artifacts that help both help people who don't agree to work together collaborate...”
  — Maya Bialik (B), 11:11

---

Important Time Stamps

| Segment | Timestamp | |---------------------------------------------------------------|-------------| | Introduction to Maya & QuestionWell | 00:30–03:31 | | AI’s impact on cognitive load and enduring learning science | 03:31–04:52 | | Addressing AI polarization and four purposes of education | 04:52–06:46 | | Patterns in AI skepticism (US vs. international, social views)| 06:47–09:15 | | Educator responses to the book; cultivating nuance | 09:15–10:25 | | Improv, AI, collaboration, and “Yes, And” approach | 10:25–12:31 |

---

Takeaways

• AI’s Transformative and Limiting Potential: AI will change much about teaching—but some “irreplaceable” aspects, especially human connection and core cognitive science, remain constant.
• Teacher Agency Must Be Central: Best practice means not replacing, but empowering the educator’s decision-making with AI.
• Complexity, Not Extremes: Progress lies in holding multiple truths, recognizing AI’s risks and opportunities vary by purpose, context, and cultural approach.
• AI as Collaborative Mediator: New tech can foster understanding and creativity when used to bridge divides—inside and outside the classroom.

---

Further Reading:
Check out Maya Bialik’s new book, How AI Changes Everything and Nothing in Teaching and Learning (Solution Tree, co-authored with Peter Nilsen).

Guests:

• Maya Bialik (Teacher, Author, CEO/Co-Founder, QuestionWell)
• Meredith Rosenberg (New Advisory)

Host:

• ASU+GSV Summit Team

---

Transcript

[A]

We are here at the ASU GSB Summit 2026 with Maya Bialik. She is a teacher, she is a researcher, she is a CEO and co founder and she is an author of the book. Her newest book is called How AI Changes Everything and Nothing in Teaching and Learning, co authored with Peter Nilsen. It's ARD now from Solution Tree. Maya, welcome to the podcast.

[B]

Thank you for having me.

[A]

You've done so much. I had read about a quarter of your bio here, but let's start by talking a little bit about Question. Well, if that's okay. So you are CEO and co founder of an AI based assessment company. Tell us what QuestionWell is, how it's grown, and how you use integration to accelerate your growth.

[B]

Absolutely. Wait, integration? Like what?

[A]

Integration with like lmss and.

[B]

Oh, okay, okay, great. I was like, it could mean anything.

[A]

It could mean anything.

[B]

Okay, so questionable started when I was teaching middle school science three years ago in 2023. And my husband is a software engineer and I'm a teacher. And I had already previously done some work on AI in education with a book in 2019, which immediately became outdated as soon as ChatGPT came out. And so my husband and I saw this opportunity where he created Questionable for me that wrote high quality questions based off of a reading aligned to learning outcomes that I could export to any of my favorite platforms. Yeah, at the time there was talk like, oh, just use ChatGPT to do it. But at that point that changes what the role of the teacher is from someone who is writing questions to someone who's copying and pasting questions, which I think is a meaningful and negative shift. And so in this case, the AI writes the questions and it helps to reason about the quality of the questions and what their coverage is and what the bloom's level is. And then the teacher can export them to their favorite platforms. And since then it has evolved, it obviously can write not only questions, but everything under the sun. That is the new table stakes of AI in education. So the question shifts. The question becomes is that are those instructional materials written by AI any good? What is the quality of them? And I mean quality in terms of like high quality and also what are the qualities of them? So there are some things that are kind of clear and straightforward. So if I ask for something to be a certain reading level, it should be at that reading level. But then there are a lot of things that are like, reasonable people may disagree. So here's a lesson plan. And someone might say it is too teacher driven. Someone might say it is too student driven. And so where it has gone is elevating teacher agency rather than eroding it, which is something that can happen if AI is just giving you instructional materials and you're expected to just take them. But we're actually building in the affordances for teachers to make informed instructional decisions in that moment of instructional decision making. And it's coming back to my roots of learning science. And so I, yeah, way back in the day, my background is cognitive science. And so at the time in 2012, I tried to make a nonprofit that was bridging research and practice gap. And at the time, I would say now that at the time it was impossible and now it might be possible. So I'm really excited about the role of AI as the sort of infrastructure that could actually bring to bear at every moment of teacher decision making. What are the relevant possible research, theories, considerations, and how to surface those trade offs in the decisions that teachers are making every day to help make those decisions more informed.

[C]

So you mentioned 2012 to today, 14 years. Students learn in a different way today because of technology and AI, availability of tools. So how has your thinking evolved now that students and how they learn has evolved?

[B]

Yeah. So that's actually a perfect segue into why we called the book irreplaceable. So everything has changed. Nothing has changed. So the core cognitive science of it has not changed. There's still cognitive load, there's still extraneous cognitive load and intrinsic cognitive load. And so the question is now, how do our tools help us teach and learn better? So in the case of cognitive load, maybe AI can help take on some of that extraneous cognitive load to help students focus on what they're learning. And maybe it can help scaffold some of the intrinsic cognitive load when they're having trouble accessing the learning. Or maybe it can help push them further, even enhancing and adding to the intrinsic cognitive load when they need to be pushed further and can take it further than they ever could before. So what remains irreplaceable? Well, among other things, one thing that remains irreplaceable is the cognitive science underneath teaching and learning. And then the question becomes, how does AI change it in order to support it more deeply?

[A]

Yeah, we're in this really interesting time where AI is very polarizing. You have people who are just all in. I don't want to call it hype, but you can call it hype, and then one could call it hype. And then there's ones who are people who are very afraid of AI. They do not want to touch it. They just feel like it's the beginning of the end. And your book really tries to break those apart and say it's not all in or all nothing. There's ways in which, like you've just named some AI can actually enhance the best practices learning, science, pedagogy based practices of teaching. Tell us about how we get out of that polarized moment.

[B]

Yeah, absolutely. That's one of my favorite things to talk about. And the way that we start every talk when we talk about the book is there are different reasons that we have education at the same time and we have multiple purposes for education and they're all kind of vying for our attention. So there's the academic, which is passing on those intellectual traditions and disciplinary ways of thinking. There's the humanist, which is human to human connection among students between teachers and students. There's the social, which is civic responsibility and transformation of society. And then there's pragmatic, which is education prepares you for the workforce and you need to have those qualifications. And all four are valid. And all four lead to complete completely different views of AI. And it's not the case that one makes you pro and one makes you anti. Every single one shows different risks and opportunities in AI. And so we have this table where we go through each one and what risks it highlights and what opportunities it highlights. And we find that that just like kind of calms everyone down because they see like, okay, my worries are represented and my things I'm excited about are represented. And there's sort of a whole landscape out there and it's not. I don't have to exist in one point at a time and be freaking out about, is this good or bad?

[A]

Right.

[B]

It is both good and bad. And it's up to us and what we do with it to make it more good than bad.

[C]

So what, what groups do you find are more afraid? So any of those that focus more on the risks versus the opportunities. Any patterns that you've seen? Yeah, patches. Either geographically, economically, type of learner, general education versus special education. Any patterns?

[B]

That's a great question.

[A]

Generation A little.

[B]

Right. I feel like I've been proven wrong on every single one of those hypotheses. I feel like it really cuts across. But I will say that one thing that's kind of underappreciated sometimes is that in the US we tend to be a little bit more negative compared to the rest of the world. The rest of the world.

[C]

The use of AI, you're saying? Because all of those four buckets. Sorry, in all of those four buckets.

[B]

Yeah. I'm not sure how it breaks down across the buckets, but I know that in the rest of the world, in terms, I guess in terms of the pragmatics, for the most part, because they're able to see it as a way to really move forward much quicker than they could before. And so what I'm hearing is that if there's any pattern that I have heard, it's about the US compared to the rest of the world.

[A]

I would hypothesize also that the people who believe that education is primarily social, as one of yours, are afraid because the way AI keeps manifesting right now, it feels a little antisocial and it started feeding into the anti social network movement at the same time, the isolating screens, all of that. So I see a lot of pushback from that camp.

[B]

Yeah. And I would say that that's also the humanist camp in a sense. So the social is like the collective. There's also like the human. It's all about the human connection. Of course, you could see AI as dismantling that. At the same time, I'm very excited about uses of AI that enhance.

[A]

I am too.

[C]

So, yeah, I mean, I'm going to jump on that third bucket though, or the last bucket about the jobs and if you. So I focus on talent and leadership. And so AI certainly is causing people to think about how is it reinventing my job and my place in the workforce. And so I find now there's so much mass media that focuses on that aspect, particularly in the US as maybe the challenges through the economy. So I could see more of the view that way.

[A]

No, it makes sense.

[B]

Yeah.

[A]

Yeah.

[B]

I mean, it is. Every new technology like has these big transformations and a lot of times the long term effects are the ones that we don't see coming that are actually the most impactful and the short term effects are the ones we're all worried about that end up working themselves out.

[C]

Right.

[A]

Yeah. As the book is out quite recently, and I'm sure people are receiving it, making sense of it, I'm sure. I'm curious what the reaction has been and what you've heard from teachers, specifically, what do you hear from educators who are at the front lines, who are dealing with all the pros and cons of this on a daily basis. How have they received it?

[B]

Yeah, I think that the way you framed it at first, I think really captures it is like it is able to hold the complexity of these kind of opposing truths all at the same time and say yes to all of them and validate all of them and even just think of them all at the same time. And so wherever people fall, when they start, they feel just like more well rounded coming out. So they don't just feel blindly more hopeful. They feel like informed more hopefully.

[A]

Right, right, right.

[B]

That's kind of what I mean.

[A]

They have more of a nuanced landscape view of the entire system where things might go right and wrong.

[B]

And sometimes, sometimes people feel like when they're more informed, they get more cynical. And so that's the main feedback I've heard is like, that's so weird. I feel more informed, but also more hopeful.

[A]

I like that. That's a good thing. Yeah. I have an odd question, but I'd love to hear your reaction to this too. One other thing that's interesting about you is you are an improv person. You're an improviser, you're a student. And I feel like one thing that's interesting about AI is it's so improvisational, it's reactive.

[B]

Yeah.

[A]

Unlike much technology where it's, you know, fill in the box, answer the question. It can react wherever you want to take AI. It can kind of follow you there. I'm curious how that informs your. Your personal use with AI, but also how you think that might inform positive uses in an educational setting.

[B]

Oh, interesting. I mean, I think it depends on how you're thinking about, like, what's core to improvisation. But actually what I was saying a minute ago, that we're able to say yes to all these different things at the same time.

[A]

Right.

[B]

And so core principle of improv is. Yes. And of course. And so thing I'm very excited about. And also tying back to when we were talking about helping collaboration, one thing I'm personally very excited about is AI being able to bridge difficult collaborations and create mediating artifacts that help both help people who don't agree to work together collaborate and come to more of an agreement. Because it's able to. Yes. And both people. Something that just happened on Slack, like literally yesterday is I posted something, or yeah, Will, my husband, who created the whole app, posted something on Slack. And the agent chimed in and gave all this. This research context that like, kind of like proved him right. And. And I chimed in and I was like, well, actually like. And then the same agent chimed in and said, I'm so sorry, and gave all this context to prove that I was right. So it was a very funny interaction, but we had all this context available to us. We were able to much easier see, synthesize everything and come up with a solution that actually addresses the problem and the vision at the same time.

[A]

That's powerful.

[C]

I do love that. As marriage counselor. Exactly. Collaborator. A marriage counselor. My husband and I also work in the same industry, not as app designer and user, but also in edtech as well too. So the debates that we have, perhaps AI could be useful that way.

[B]

Absolutely. I'll recommend you to my AI.

[C]

I'm your agent.

[A]

Thank you so much. Everybody has to check out How AI changes Everything and Nothing in Teaching and Learning by Maya Bialik. We're here with Meredith Rosenberg from New Advisory. Maya Bialik, teacher, author, you know, and founder. Yeah. Multi hyphenate human being and human being at heart. Thank you for being here with US at the ASU GSB Summit 2026.

[B]

Thank you.

[C]

Thank you.