Cognitive Surrender: What AI Is Costing Our Students

Matthew Pietrafetta and Andrew Ferguson discuss cognitive surrender: what students give up when AI does the thinking, where AI helps a tutor, and how families can help kids decide when to use it.
Beyond the Test · Season 1, Episode 3 · 36 min · Listen on Apple Podcasts · Listen on Spotify · RSS feed

About this episode

Most students now use generative AI in their schoolwork, which makes the question for families and educators less “will they use it?” and more “what do they give up when they do?” In this episode, Academic Approach founder Matthew Pietrafetta and Andrew Ferguson (on Andrew’s birthday) dig into what Matthew calls cognitive surrender: handing the hard thinking, planning and evaluating over to a chatbot.

They talk about why productive struggle matters for adolescent development, where AI genuinely helps an expert tutor and where it falls short, what the six Stanford criteria for high-quality tutoring say about expert materials and human relationships, and how schools are already responding. They close with a practical framework for families: help students decide, on purpose, when to outsource a task and when to do the work themselves.

Key takeaways

  • Cognitive surrender (also called cognitive offboarding or outsourcing) is the term MIT’s Ad Hoc Committee on AI Use in Teaching, Learning, and Research Training used in its August 2026 report on AI and education for students falling back on AI at the first hint of struggle and giving up the critical-thinking process. Matthew also points to EEG research from the MIT Media Lab (Your Brain on ChatGPT, a 2025 preprint) that found weaker brain connectivity when students used ChatGPT to write than when they wrote on their own.
  • The cost goes beyond a single assignment. Productive struggle builds planning, frustration tolerance, delayed gratification, self-management and the habit of asking adults for help. Skipping the struggle also skips the ownership: if ChatGPT writes the essay, there is no “sweat equity” in it.
  • Not all outsourcing is the same. Matthew outsourced directions to Waze without a second thought, but would not outsource writing a paper. The goal is an informed choice about what to hand off and what to keep.
  • Students are losing the planning step, not just the answer. Andrew’s version of the Mike Tyson line: students used to avoid getting punched in the mouth, and now they are avoiding making the plan at all.
  • AI can help a tutor, for example by finding the calculator steps for a graphing problem or generating extra practice questions, but the tutor needs expert knowledge and a real relationship with the student to evaluate what comes back. Most students do not yet have the evaluation skills to do that.
  • The goal of tutoring is independence, not dependence. Matthew describes his job as writing himself out of a job, using question-and-answer dialogue that builds the student’s own internal voice. A chatbot that simply hands over the answer does the opposite.
  • Expect schools to push back (blue books, in-class writing, restricted devices) and expect more research on where human intelligence and artificial intelligence each belong in education.
  • The family takeaway: since AI is not going away, help students become informed consumers who understand what they give up each time they surrender a thinking task.

In this episode

  • Why cognitive surrender is Matthew’s biggest concern about AI in education
  • Productive struggle, and what students lose without it
  • Age limits, regulation, and why the genie is already out of the bottle
  • Who you want to become: ownership, effort and “nothing worth having comes easily”
  • Deciding what to outsource: the Waze example
  • Planning, executive functioning and getting “punched in the mouth” by a hard problem
  • How tutors can and cannot use AI: curated materials, relationships and evaluating output
  • Why the goal of tutoring is independence
  • Where AI in education is headed over the next five years
  • What families should do now

Who’s talking

Matthew Pietrafetta (Founder & CEO) and Andrew Ferguson (Director of Business Development) host this conversation.

Keep reading

Full transcript

Read the full transcript (6,401 words)

Lightly edited for readability. Beyond the Test is a conversation, so please check the audio for exact wording.

Matthew Pietrafetta: All right, hello everyone. I’m Matthew Pietrafetta. Welcome to Beyond the Test, the Academic Approach podcast. We have a great topic today. We’re gonna talk about AI, artificial intelligence in education. And I’m here on an auspicious occasion with my esteemed colleague Andrew Ferguson on his birthday. So happy birthday to Andrew. We’ve been working together for, you know, over twenty years and since you’re turning 29 now, that means I guess we started when you were nine, right?

Andrew Ferguson: Thank you. More than half my life, yeah. I was a wunderkind. I don’t remember a minute of it, when you’re nine, it’s hard to keep all those memories together.

Matthew Pietrafetta: Remember when we started that middle school tutoring program when we were in, like, fifth grade, fourth grade? We were teaching those first graders. We were great. Speaking of nine-year-olds, they’re now using, not our tutoring that we provided for them years ago, but perhaps ChatGPT to answer some of their questions. So that’s some what we’re gonna be talking about today and how we feel about that. What guidance we have for, you know, educators and parents around that engagement with AI.

Andrew Ferguson: So with all that in mind, I have a few questions for you and I thought that could kind of start the conversation up and then we could also remember at the very end, we’re gonna talk about some things that families can take with them after the podcast. So let’s roll. So I thought about this topic because if you remember a year ago, you and I, as well as Kyle Benjamin, our esteemed colleague, we were at the NACAC 2025 conference, and the big theme, as you remember, was all things AI. AI was going to be a wonderful solution to so many people’s problems in the college admissions process. And so it got me thinking about AI in general and how it affects our lives as educators, as tutors. So the first question is a fun one. I want to know, in terms of AI and tutoring or AI and education, what is your absolute hottest take? If you’re kicking the door in on this topic, where would you start?

Matthew Pietrafetta: I start where we spend all our time, where we’ve spent our time for, you know, for many decades now, which is in the student learning space, like the K through twelve developmental journey the students go through in learning and developing critical thinking skills, executive functioning skills, resilience, perseverance, flexibility, problem solving, self-monitoring, self-management skills, all those skills that we realize are part of the journey that we support as tutors, and that we always want to kind of impact on multiple levels, skills, you know, non-cognitive factors, all the things that make someone a great problem solver and a successful student, and then a successful professional in life. And and so from that perspective, my hottest take is I worry about what in a recent MIT study they refer to as cognitive surrender. It’s also been called like cognitive offboarding. Or you know what I call cognitive outsourcing, which is basically giving up on the critical thinking process and going to ChatGPT to answer all your questions. And just to unpack that, why I think that’s my hottest take and my biggest concern is that that process of problem solving, right? That process of critical thinking that involves going up Bloom’s taxonomy all the way to the level of synthesizing and evaluating and making judgments. You know, developing knowledge, retrieving that knowledge, applying that knowledge, evaluating critically and making decisions and then generating new things. That journey of learning it exercises the brain. And as the research has shown, and when they’ve hooked up, you know, brain scanning technologies, students using ChatGPT when they write their papers versus those who write them authentically, the brain is simply dormant when AI is doing it for you, and the brain is incredibly active when you are doing it yourself. So when you cognitively surrender to AI, when you outsource, when you offboard these cognitive tasks, your brain is not exercising. In addition to that, you’re not developing, right? A lot of those ancillary skills are so important, like the ability to deal with frustration, tolerance, right? When something’s hard, right? When you surrender and you don’t go through that. That frustration, or that productive struggle, as it’s called in education, then you’re not developing these other skills like planning and deferring gratification and self-management of emotions, right? And seeking advocating for yourself and seeking help from adults in office hours or through a tutor in a conversation. So all these other ancillary skills and you can enumerate them and when you put them on a list, you’re like, my gosh, all those things are critically important, I think, for human development, right? Especially adolescent development. That is all being surrendered, right? When you pop into AI for the quick fix, for the immediate gratification. Right. So it’s that cognitive surrender and the developmental consequences of that, frankly, during a critical time of adolescent development, that’s my biggest concern.

Andrew Ferguson: Before this podcast, I was trying to find something on whether or not there was any conversation about having, have to be a certain age, like 18 or over to enter into, and there’s a lot out there, but it’s mainly just parent groups advocating for wanting it. But there’s probably just as many who are advocating for being 18 years of age to access Instagram, and that’s never happened. So it’s, but, I mean, obviously this is different. Just, feel like, yeah, it’s all the things that you just mentioned, you know, you’re the brains of this operation. You know those topics at a much higher level, but I also think about how all of those things lead to your, who you become, right? Like how you learn, you know, the struggle of having to learn independently. It crafts who a child is going to become as a young adult and inevitably as an adult of like the work required to, you know, to accumulate knowledge. And so I think there is, I think there is something to be said for, I’m not, I am not one to put restrictions on people in terms of this, that or the other, but it, if it were me, I would want to know if my child was really gathering this information. And I feel like, There were, like I said, lots of conversations, the wonderful chasm that is Reddit, lots of parents in there talking about their want to make sure that there has to be some sort of age requirement because some of the things that are being talked about on, that some of the, some of the advice that ChatGPT and other places like that or other AI platforms are providing isn’t necessarily built for the eight-, nine-, 10-, 12-year-old, whatever. Yeah, I think that’s my hot take is that there may be a point where not all advice or not all answers to questions should be answered at certain points in life. And perhaps that’s the job of the parent to be monitoring that, but also I think it’s a world in which it’s gonna be so abundant and constantly around children and in the learning environment specifically, but yeah.

Matthew Pietrafetta: Well, I think that outcry for regulation is far and wide, both from parents and from corporations, around privacy and security of data and so forth. So the but it is it’s on a certain level, you’re so right, like the genie’s already out of the bottle. And so like yes, there’ll be this progressive attempt to kind of rebottle the genie, but we know how hard that is. And so I like that point you made earlier about, you know, who’s the person you want to become? And what is that what does that entail? And I’ve had mentors say to me in one form or another over the years, and I pass it on to my students and my children, like nothing worth having comes easily. Right. And if you think about if you think about what that means, it’s like if there wasn’t some productive struggle in there for you to earn the outcome, right? To the sweat equity that goes into something, then you don’t value it as much, right? Because there’s no Like there’s no that ownership bias that we all have when we make something, we’re like, I made it, so therefore it’s awesome. Right? You don’t have that when ChatGPT hands that essay over to you. There’s no ownership there. There’s no process of development that you can claim some sweat equity in and some work that you put into it to make it of value to you. And so who are you? Well, you’re not the you’re not the you are not the author, you are not the creator, right? You can’t claim that. And what’s more, you sacrificed a journey, right? And a process of journeying through you know, to achieve something, which so many of our students know who are play a musical instrument or play a sport or whatever, they know what it’s like to go you know, go to coach and get feedback and challenge themselves and improve and play through some adversity and things like that, and that feeling of accomplishment they have. That is always we’ve always felt you and I as coaches that we parallel that in the academic space, right? With the same kind of stick to activity and perseverance and setbacks from adversity, but coming back and learning from it, self-reflection, self-direction, and you go on. Well, all that all who do you want to be? Who do you want to become? Well, you want to be become someone that has those values, right? That kind of work ethic, because we believe there’s something transcendent about that beyond the a that you get or the great standardized testing score that you get, you become someone who understands how to sacrifice, work hard, self-manage, self-reflect, self-correct, and improve. And so that’s what’s kind of being given up. And I guess to the point of like regulations, I and this is where I kind of move in this conversation to parents and educators, the genie’s out the bottle. So you’re not going to be able to regulate it like anytime soon as far as I can see. So how do you then, if you can’t regulate it, how do you mediate it? And by that I mean like how do you Create constructive conversations for students so they understand what’s at stake in using it. Like, do you understand, even and I think most students don’t, that when they cognitively surrender and go to ChatGPT, that they’re getting immediate gratification. But what are they sacrificing? Like I don’t know that how many conversations they’ve had with anyone about what’s happening to their brains and what isn’t happening to their brains, what’s happening to your point to their character. To their development, right? Like the person you want to be when you go there and you instead of going over here and going on this path. So I think it’s like since you can’t put the genie back in the bottle and maybe you can’t, you it regulations won’t come in time enough to kind of save a generation. How do responsible and well informed adults and educators frame a conversation for students where they become more informed consumers? And they start asking, they understand the consequence of going to it, what they’re giving up when they choose to outsource or surrender, and they make critical decisions about when they will and when they won’t. Like I’ll give you an example. This morning, going to the gym, a friend of mine, his car’s in the shop, so I went to pick him up and he texted me his address. And I did not go and get a map and like unfold a map, and go look where his street was in the town that he lives in, which I used to do like decades ago. You know, growing up as a driver in Boston to figure out where I was going to meet my friends. I didn’t do that. I outsourced it to Waze. And Waze did a brilliant job and got me there in no time. And I would I knew I was outsourcing a cognitive task, but the efficiency was great. And I didn’t feel like I really have to learn a map of the town that he lives in order to become a better human being. So I chose to outsource that to artificial narrow intelligence, ANI, which is what like Waze is, or what ChatGPT is. It’s not super intelligence or general intelligence. It’s a narrow intelligence, it’s part of our brain that we could do ourselves, but we can outsource now. And I chose to do that, but cause the convenience outweighed, I thought, the learning opportunity, honestly. But I think those are the types of conversations students should be having with themselves. And I think they’re not. They’re just going to artificial narrative intelligence because it’s there, it’s immediate, it’s gratifying, but not understanding that there’s a right and a left path to take and there’s a consequence of taking one versus the other. And that’s I think if we can cultivate those conversations and that kind of awareness of what’s at stake, I think that’s probably one of the best things we can do to make them better consumers going forward because we cannot bottle the genie, right? We can’t.

Andrew Ferguson: I was thinking about, you were speaking so eloquently about the MIT study about, you’ve got at your fingertips in your brain, you’ve got that Bloom’s taxonomy. And I will take my version of that and say that the great Mike Tyson once said that everybody has a plan until they get punched in the mouth. And I think that I used to think the kids would be are always trying to figure out a way to avoid getting punched in the mouth. But it seems like now they’re trying to avoid making the plan. Does that make sense? They’re like they’re fearing. They used to fear the outcome. Right. And they I don’t want to get a B or I don’t want to get a C or I don’t want to do poorly or whatever it may be. And now they don’t even want to do the plan part. And I fear that. Some of my fondest memories of high school or college or whatever it may be was all the work that was put in. Yeah, you get a good grade or you do well on this or that or whatever, you only feel good about it because of all that time and effort that went in. I don’t understand the payoff if there was no investment.

Matthew Pietrafetta: And I’ll add to that point, I love the Mike Tyson line and it’s I think the I think that, you know, as people who spend their days and nights in the problem solving process of like highly rigorous standardized testing, you know, items that are presented to students, they punch you in the mouth. A difficult problem punches you in the mouth, and that is to say, like you go to solve it and it resists, it fights back. You try something and it fails. It literally fails. And you’re like, okay, well, what do I do now? Panic? Retreat? No, hesitate to try anything else? No, we know what it is. You go to plan B, plan C, plan D, and that’s what we do. We teach these multiple approaches to problem solving. Implicit in that is the assumption that you’re gonna get punched in the mouth, that you your first attempt will not succeed on a highly rigorous item. And so you prepare for it.

Andrew Ferguson: The part they’re lacking is thinking about the next time you have to do this. You didn’t learn anything, right? Like you made that plan and there’s no mental note of, I… I told myself I was going to study or I told myself I was going to do this, write that paper over the weekend and I left it for the night before and I have to adjust. There’s no plan. It’s just if it’s being, if too much of it is being done for you, well, that’s for a different day.

Matthew Pietrafetta: I mean yeah, I mean you’ve basically replaced, you know, what the what they you know the executive functioning skills, one of which is this planning and organizing and calendaring and deferring and time management and self monitoring that goes along with self-management. Like all that has been replaced by the immediate gratification of an answer that’s given to you. And so to your point then you don’t have you haven’t developed the executive functioning skills to know how to plan and execute on a process of problem solving. So the next time another project comes up, you haven’t figured out well how to lay how do I calendar that better. Last time it didn’t go so well. I was up till, you know, midnight the night before. Could should I start three days earlier? So that learning process of all the executive functioning, so that’s not taking place. That’s been surrendered as well. So it’s back to that cognitive surrender. What are you are surrendering brain function when you go to AI because AI is essentially artificial narrow intelligence is pretty amazing. Because it does what our brain it does a set of functions that our brains can do on their own. But like we’ve all been, we’ve all been wowed by ChatGPT. It does it way faster, right? Does it way faster. It’s so speedy and intoxicatingly immediate that that you can’t help but marvel at it. But it and so you know what you’re doing when you are surrendering is you’re surrendering that particular set of cognitive functions to this tool. And you’re not doing them, right? Like I said, with finding my friend this morning, you know, finding his address to pick him up, to go to the gym, I was cool with letting Waze do it for me.

Andrew Ferguson: I never saw you as a Waze guy. I always figured you’d be a straight-to-Google-Maps or an Apple Maps guy, but Waze keeps the police off your back, I’ll tell you. But yeah, that actually leads me to a second question, with all this talk of where AI is entering our lives in a negative way.

Matthew Pietrafetta: Yeah, you’re picturing me as like more of a boomer than I am, you know.

Andrew Ferguson: Have you, the tutor, the instructor, ever used AI as a helpful tool for your lessons?

Matthew Pietrafetta: Yeah, that’s interesting. So just last night I was working with a student and she had a TI-Nspire calculator, which a lot of our students have exposure to, but most of us are using TI-84s or TI-89s or whatnot. So the Nspire has a few functions in there that like she was trying to reset rectangular coordinates within her graph. And I know how to do that on a bunch of calculators, but not as well on the Nspire. And so I would historically go and Google that, but then get a bunch of like links that are not useful for me, right? It’s just like an inefficient search. But you go into ChatGPT, you ask about that, and I got the three steps that allow her to reset her graph immediately. This search function is amazing for information. I’ll use it for a variety of like Google search what I used to use for Google search to find information, much faster. It just curates information more effectively without all the ads and the interruptions and distractions of a typical Google search. So yeah. But you know, the National Student Support Accelerator at Stanford that approved us for our tutoring model because we met all six criteria of it you know a few years ago

Andrew Ferguson: The cool badge in our email signatures.

Matthew Pietrafetta: The badge and our signatures. Yes, exactly. So we got to prove for our you know high impact, high quality in instructional model that met those criteria. Two of those criteria are really important. And one is well curated instructional materials, like expert materials, right? And I do not trust what I get from ChatGPT when it comes to that. And I’ve seen students use it to get explanations that I think are either wrong or incredibly inefficient. Like I would never have you do that. Or think about it that way if you’re trying to finish a standardized test on time with accuracy, right? And so don’t trust the materials. Whereas we’ve you know been curating our materials for over 25 years with a lot of knowledge of the test. So I’m not going to get my materials from there. And Stanford just updated their study when it comes to AI and this model. And there’s uncertainty around the quality of the instructional materials, right? That’s one of the six criteria. And most importantly, the relationship between the adult educator and the student. And that relational context to personalize, to connect, to contextualize, knowing the way the student learns, knowing what the student needs, knowing the history of the student’s performance and the subtleties of the student’s personality, right? Like that. Relational context for instruction is one of the six criteria that AI cannot meet. And so for me, like I, you know, I think that the the as far as a supplement to what I do, I don’t need it. And it doesn’t answer the questions. And the things good instruction can provide, which is the expert knowledge of what materials are the most effective. And have been research based and proven to be effective. And then the adult educator with the right supportive, motivating relationship with the student, those elements, you know, AI cannot produce with with any kind of validation at this point in any of the studies that have been done.

Andrew Ferguson: Yeah, I think independently, no, it can’t. I do use it to provide extra examples and extra problems, but I guess I follow the path that they would use in the sense that I’m doing it based on where I feel like I’m just looking for additional examples of questions that we can work through or if we’ve already, if they’re still struggling after having done the material, the examples we’ve provided, I can look for extras. But with the idea behind that, I know what the student’s needing. It’s not just blind.

Matthew Pietrafetta: That’s it, yeah. That’s the two things: you’ve got a strong relational context with the student. You know the student, you develop personalized relationship, you know exactly what you’re looking for, and then you know what good material looks like. So if you can then make an evaluation, and this is the problem, right? Like in Bloom’s taxonomy as you move up. You want to be able to analyze and evaluate information. Well, evaluating means that you have the critical judgment to know if it’s good or bad, right? If it’s correct or incorrect. And so you know the material really well. So then you can evaluate what’s coming out of ChatGPT for you. And you know what you want to give to the students. So you’re personalizing the content for them. So that makes sense to me within the traditional structure of what really good instruction is. But here’s the concern, and this is what we’ve seen with our students. They don’t have the skills to evaluate yet. So what gets spit out to them from ChatGPT, they cannot do what you do with it, which is say, no, this is not what I’m looking for. This needs to be modified. They’re just going to accept it at face value until they’ve developed those evaluation skills. However, if you’re cognitively surrendering the process of marching up Bloom’s taxonomy and all the productive struggle along the way of getting there, then you’re not developing the evaluation skills. So then now we’re producing possibly a generation of future AI users that don’t know how to evaluate the content that comes out of it. And we need those critical thinkers because we are going to be continuing to use it, but we need strong evaluators to determine, you know, good, bad, or indifferent.

Andrew Ferguson: Yeah, it does kind of upset me when I can tell. I don’t know if you’ve been in this situation. I see it most in academic support. Don’t think a lot of the students are cheating themselves for the ACT and SAT in terms of like cheating themselves out of learning the material. I do see, I do see some times where a student has reached a certain point in it, whether it’s calculus or chemistry or whatever, whatever math class I’m helping them with and where you, you can tell that they have, they’ve had far too much support at this point to be able to work independently and that, and you know. In that session where the test is tomorrow, you begin to worry of like how, when you take away the support mechanism that’s providing 50, 60, 70 % of the knowledge, you’re gonna get that punch in the mouth that you were hoping to avoid.

Matthew Pietrafetta: Yeah. I’ve always said to students, yeah, I’m with you. I’ve always said to students, my job is to kind of write myself out of a job. That is to say, the dialogue that we have about problem solving, the kind of modeling and scaffolding that I do, you internalize it on your own ultimately, and you don’t need me anymore. Because you can kind of talk yourself through that problem solving. You can function independently to your point. Like that’s the goal. The goal is not dependence, the goal of tutoring is independence. And that’s a great way to frame it: our ethical responsibility has always been, and we’ve talked about this in the company for a long time, not to be handholders or like crutches, right? Because if you’re cultivating dependence, you’re not setting the student up to ultimately go into the test, which they have to take on their own and be successful, and then go on in academics and be successful on their own. So we’ve always thought a lot about how do we make sure our instruction does not cultivate dependence, but rather cultivates independence. And that’s really intentional work because you have to ask a lot of questions. You have to make sure the student has internalized how to ask those questions, how to have the internal dialogue with themselves when they’re taking the test and talk them through it with some of the scripts that we’ve taught them. And that’s precisely what is missing. And you know, if you think of the way that and we’ve done this before, I’ve done this in presentations on AI before and talked to you, you know, we’ve talked about it, which is, if you put a math problem into ChatGPT and see how it responds with the answer, side by side with what a question-and-answer dialogue would look like between you and your student, the amount of work that we’re asking of the student is so very different from what ChatGPT does, which is basically they’ve surrendered all the work to ChatGPT, they’ve outsourced it all. And instead, you know, what you’re doing when you walk them through that instructional dialogue. Is building their independence and their own critical thinking, which then they can internalize, which then they can do on their own, which then they don’t need us for anymore one day, as opposed to the complete surrender of that entire process. And it’s really interesting, kind of side by side that I did it in a presentation once of like, here’s a question we’d ask, student would answer this, we’d ask this follow-up. And you go down and it’s you know a hefty dialogue, and then here’s how students use ChatGPT. Solve this for me. And then it spits out something. And so that that’s that’s really the critical difference that work has been outsourced, surrendered, and we don’t want that work surrendered because we believe that there’s some transcendent value to that work, you know, for the students’ development.

Andrew Ferguson: Right. This time of year is an interesting time where juniors are just starting this process or have been starting it for the last couple of months and seniors, you know, they’ve got this one last test here in the next three days. And it’s funny how when you have a lesson with someone you’ve been working with for a while, all you have to do is just give them one or two words of the rule. They miss a question that you know they shouldn’t have missed, and they’re gonna know they shouldn’t have missed it. All you have to do is nudge them in the right direction because it’s a learning process, but it feels like when it’s AI driven, every time is the first time because they just never had that magical moment of learning. They were just duplicating whatever was fed to them. So, yeah. We have one last question before we get into some helpful hints. And that last question is, where do you see this going? Like where do you see AI in education, and let’s start with tutoring, in the next five years?

Matthew Pietrafetta: Yeah. Well, I think like I’ll answer that in two ways. One, there’s immediate reactions going on, you know, you look at on the university level, like moving back to blue books, even high school level, teachers going back to in-class writing, right? University of Chicago, like one device in the classroom, one note taker shares the notes out with everyone else, but no one else can have devices open. So there’s gonna be a reaction against the genie being out of the bottle to try to bottle it by basically restricting it. And that’s good because then you can get more authentic understanding of what your students know and don’t know, which is a huge concern. So that will we’ll see a lot of that. Then we’re going to also see similarly just a bunch of research going on, like the stuff that just came out of Stanford, of saying, like, well, let’s talk about it. How do you use AI in productive ways in the classroom? And I think like your example was really great. There’ll be more research on that. Take a great instructor like you who builds amazing relationships with your students, who’s who knows everything about the scope and sequence of high school mathematics that you need to know in order to be at a expert level, college ready level. And put some AI tools in your hand and study it. And you’re already using it a little bit, finding those items, creating those items that you know are going to personalize the experience for your student. But under the right guidance and guidelines, how can AI be a support? Right. And I think that there’ll be a lot of research on that, of like within what boundaries. And I think what we’ll find is this really interesting. My suspicion is we’re going to find, and it’s going to be really important for humans, that we’re going to have a column of human intelligence and a column for artificial intelligence and education. And we’re just gonna put down the criteria or the activities or the functions of where human intelligence is really important. And it’s gonna be unsurprising. Empathy, perspective taking, relationship building, right? Like those things are transcendent and timeless, and they will be forever and they will not be replaced. Right, by robots. And so that and we’ll know squarely, okay, well, let’s not try to mess with that. Those are the things that we value. Let’s make sure humans really can relate to humans really well. And then we train teachers to do that really well. Because all the research shows that when you have a mentor that understands you, if you have a doctor who cares for you, you do better in the hospital, right? If you have if you have a teacher, an adult educator, a mentor, or a coach who cares for you, you’re gonna have higher self-esteem. You’re gonna be less prone to all sorts of negative behaviors, you’re gonna have better outcomes, right? This is all proven. So, and then in the column of AI, we’ll say, well, that that’s a lot more efficient if the robots do that. And why should we be bothered with that that administrivia, that kind of like administrative trivial work right over here? So let’s just offboard that, let’s outsource that. It’s kind of like me with Waze this morning, right? And we’ll feel okay about it. But I think that’s what’s missing: there’s no clarity right now about where the human intelligence column is and where the artificial intelligence column is. They’re separate and you don’t cross. You enumerate those tasks separately within those functions and you’re pretty clear about it. So I think that research needs to be done. And then I think that clarity will come. And it probably won’t be terribly surprising what lives over in the human column and that we hold as sacred and that we think is really important and precious to us as an evolved you know primate species. And then what’s over here for the robots to help us with. And I think that that will be there’ll be a lot of research studies around that.

Andrew Ferguson: Those robots would be really busy if they’re planning on taking over the world in five years. They have other things to do. Stop ruining children’s learning. So if you were to give families a takeaway, what should they be thinking about? In terms of their child’s education, partnered with the world of AI that is unavoidable. What would your big takeaway be?

Matthew Pietrafetta: Well, I think the role of the parent and the educator, again, back to that theme: We want to help our students be independent. And that means that they need to evaluate and make judgments and self-monitor and self-manage and self-regulate themselves. So to give them a framework and encourage them to use a framework for decision making about when they use it and when they don’t. And I that that goes back to my earlier point, which is I don’t think that students understand. When they cognitively surrender these cognitive tasks, these intellectual tasks of planning, organizing, as you were saying earlier, evaluating, synthesizing, arguing, right? Like when you surrender that, what brain activity have you surrendered? What cost is that for you developmentally, in terms of your long term skill development, in terms of your emotional regulation and development, in terms of your own self-confidence, in terms of, like you said earlier, who you are as a human, right? Is that something that you want to outsource? Well, picking up my friend at the gym this morning and figuring out where he lives? Okay, I’ll outsource that. That’s just convenient. I don’t want to go get a map somewhere. I don’t even know where to get a map these days. So I’ll outsource that. Fine. But I don’t want to outsource writing this paper because I really do think that if I force myself to read through this novel and reread it and synthesize and analyze, that that’s a task that’s developing my brain to critically think. To evaluate, to synthesize information, to problem solve, to reconcile maybe apparently disparate things and find some way to unify them, right? Which we seem to have a lot of trouble with in politics these days. People finding common ground and synthesizing and unifying. So like maybe there’s a bunch of other ancillary tasks I’m learning here that I don’t want to that I do not want to outsource, that I do not want to surrender. And so we have to educate them: when you do surrender, this is the brain development, the brain experience, the intellectual development that you are not enjoying, that you are not participating in. Do you want to do that? So long as you become aware that when you go down that left path towards AI, you give up this right path towards human intelligence and human development, okay, fine. Now you’re the decision maker. What do you want to do? And I think we need to make them informed consumers, because right now they are not, as far as I can tell. And I think that’s the best thing we can do. And then we get we have to leave it up to that generation of students to navigate that decision. But first it has to be framed for them. And I think the genie got out of the bottle so fast that there was no real critical thinking about when I go left down that AI path to get all my answers, what does that mean? And I don’t think we’ve framed it to explain what the right path promises, what it has meant. Over the past hundred years of education in K through twelve and and what it could mean for your future. So I think we need to frame that conversation, including how to get back to that path.

Andrew Ferguson: And how to get back to the other path once you’ve become so reliant on it. Matthew, this was fun. And thank you. Thank you for the birthday podcast. I appreciate it. And that was great. We will, we will come back soon with episode four. Very, very soon. But Matthew, any parting words?

Matthew Pietrafetta: Yeah, just let’s celebrate Andrew’s human intelligence because it’s prodigious and means a lot to a lot of students, a lot of alum and here’s to another year of it growing and developing and not being replaced by AI. All right, take care.

Andrew Ferguson: Thank you. Have a great afternoon. Thanks.

Are you ready to start tutoring?

Start with a diagnostic test or receive personalized guidance on where to start. 

Share this post:

Check your inbox! PreACT scores returning for Chicago Public School students.

Vanderbilt Reinstates ACT/SAT Requirement for Fall 2029 Applicants

Hey — we've seen you around.

That tells us you’re serious about your student’s future.

So are we. Let’s stop circling and start a real conversation.

Chicago Public High School Applications open now!