🎥Webinar recording


Nina Bamberg (00:00):

Welcome to today’s webinar on empathy in the Age of AI. We are super excited to be joined today by our friends at Thinker Analytix, and we will introduce ourselves in a little while or in the next couple slides. And we are really excited to have this conversation today. And just a general outline of what we’re doing today or what we’re going to talk about. We’re going to start off with just a quick discussion of why empathy, why are we talking about empathy specifically in the age of AI, why it’s an important skill, and the types of context where students will really need this skill of empathy, specifically because of AI technology. And then Aidan is going to give a presentation on systematic empathy and building that skill in the classroom. And then we’re going to close out with a real life example from a video we found where students might encounter the need for this skill in real life or a real life example of an argument that was had about AI. And we’ll break it down together. And then just a quick wrapping up with some examples of using AI to build some of the argumentation and empathy skills that we’re going to talk about today.

But first, a little bit about who we are. My name is Nina. I’m the Director of Growth here at pedagog.ai, and we’re an education technology company, and we focus on providing educators with resources, tools, knowledge to navigate the world of AI and make sure that it is not a scary technology or that is not something that needs to be feared or totally avoided or even completely embraced to the point of excess. But we’ll talk a lot about that a little bit more today, what that means. I don’t know, and I see you’re here. I don’t know if you wanted to jump in at all for the intro. Our CEO Priten Shah is also here. He might jump in.

Priten Shah (02:28):

Hi. No, I think you covered it all. I think we try to lead the important conversations on AI and education and try to take a balanced viewpoint between we’re really excited about the potential that it offers, but I think we’re also cautious about the harms that can cause our students if we don’t approach it appropriately. Hopefully today we can talk about one of the ways in which think what appropriate means, I think. Yeah.

Nina Bamberg (02:57):

Awesome. Yeah. All right, Aidan, I’ll pass it over to you to give an intro on ta.

Aidan Kestigian (03:02):

Yeah. Well, first of all, thank you, Nina, for the invitation. Yeah. My name’s Aidan Kestigian. I am the lead of curriculum at this little nonprofit called Thinker Analytix. I’ll tell you a little bit more in a bit about what we do and why we do it, but largely we’re trying to help learning and working communities have better and more productive disagreements. We think disagreement is a pretty fraught topic right now in educational spaces, especially before I worked at TA. I spent 10 years in the college classroom, which is a particularly fraught space right now. But we also think that disagreements can be really generative and productive and a really important place for learning. So that’s why we’re here to talk today about some tools, tools that you can use to have better, more productive disagreements. And we’ll dive into that in a little bit.

Nina Bamberg (04:06):

Perfect. So as I said to start out, the first thing we’re going to talk about is just why empathy. Why is this such an important skill to have as a result of the advent of new AI technologies that we’re seeing today? And we see a couple reasons why it’s going to be so important. One, because we’re going to be getting a lot of information from AI, and we are going to need the skills to thoughtfully analyze that information. If we’re getting arguments from AI, how to dissect them, how to fully understand them, we’re going to need those skills. And then possibly the biggest reason we’re going to need this skill and the age of AI is because there are a whole list, and we’ll go over a little on the next slide of ethical and moral discussions that we’re going to need to have because of AI technology and different people are going to have different opinions, different thoughts on all of those issues.

And we need the skills to listen to each other and understand each other in order to navigate some of these really critical questions that we’re going to be dealing with as AI further impacts our society. And here is just a short list of the kinds of ethical and moral discussions that are really going to be coming up because of AI. So things like how it’s going to affect the workplace, asking questions about whether government should regulate AI in order to prevent job displacement, how it affects different industries. So here we have the examples of policing and healthcare. So asking the big questions about if predictive policing is ethical, if using AI to diagnose and prevent diseases is ethical. And then the next couple examples of advertising and deep fakes, those are ones that we might encounter on social media. Is it ethical on for personalized targeted ads or are they good and convenient for people?

Are DeepFakes ever ethical? Is the use of someone’s likeness or creating a video that looks like a well-known person? Is that ever something that’s okay to do? Right? Then we have some other technologies like facial recognition. What are the ethical ways that that can be employed, if any, how? Again, the justice example goes back to maybe the policing one, but can it be used to ever ethically be used to predict criminal behavior? And then the next example of self-driving cars. We’ve already seen some of the dangers of self-driving cars and the kinds of questions that people making those technologies are grappling with, as well as privacy, what level of privacy can we expect in the age of AI? And so these are just some of the really important questions we see our students needing to be prepared to engage in as AI further impacts our society. And the ability to have good productive disagreements is critical at this time.

And just quickly, just some of the places that they’re going to be encountering some of these arguments, right? We’re going to be seeing them coming from political leaders, the AI developers who are developing the technology. We’re going to be hearing from people like ethicists and academics, and then from everyday people on social media or even in person, in discussions with your friends and your family, these things are going to come up and disagreements are inevitable. And so some of the skills that students are really going to need are those critical thinking, evaluation, listening, research type skills in order to effectively engage in these discussions in a world with AI. Okay, so now I’m going to pass it off to Aiden, who is going to talk to us about systematic empathy and building that skill.

Aidan Kestigian (08:40):

Awesome. Thank you. Yeah, I mean, Nina just gave you two really important reasons why skills like empathy are critical, especially for tackling topics surrounding ai. And I’ll add one more reason to the pile. I’ll drop an article in the chat that I was noodling on over the weekend. There’s this growing body of articles. I feel like this article gets printed every two weeks now about basically how the admin and the increase of use of AI in the workforce is actually going to mean that human skills, skills like empathy, skills like listening, collaboration, are going to be even more important than they are now, if that’s possible. One of the quotes I found particularly poignant was that our abilities to effectively communicate, develop empathy and think critically have allowed humans to collaborate, innovate, and adapt for millenniums. Those skills are the ones we all possess and can improve.

I would question if we all possess them to a relevantly good degree, yet they’ve never been properly valued in our economy or prioritized in education and training that needs to change. That’s exactly what motivates our work at Thinker Analytix. And I’ll show you some of the tools we’re using to try to address that challenge. So let me drop, let me see. So I’ll drop these slides in the chat as well, so everyone can have a copy because there’s some links throughout to resources and things like that. So let me drop this deck in the chat and then I’m going to screen share. Let’s see if I can get this at the same time. Sorry, everybody’s watching me. Here we go.

All right, so let’s talk a little bit about systematic empathy. Before we start, I’ll just give you a sense of where I’m coming from at this problem about empathy and about tackling arguments head on productively. As I mentioned, I work at a nonprofit called Thinker Analytix, which is, as I said, a nonprofit originally spun out of the department of Philosophy at Harvard. Really, what we’re trying to do is give educators and students tools really practical tools for clear communication in the face of disagreement. I’ll focus on one of those tools today and then I’ll point you to some other resources where you can learn more about our work in this area. But the objects, the sort of things, the stuff of our work, the things that we focus on, the kinds of communication that we focus on are arguments. You all know actually, you know what arguments are, but I’ll just give a definition just so we have a shared terminology.

When I say the word argument, I mean a form of communication that uses reasons to advocate for a belief. You might call that belief, a thesis, a conclusion. We call it a main claim, a thinker, analytics. But that’s all you need to make an argument, at least one reason and the thing the reason is supposed to support. We think arguments are really powerful. They’re powerful for persuasive goals, but also arguments are really powerful, especially for students who are still formulating their own ideas about what they think about big issues, right? By working through arguments, you can explore many different reasons you might hold, maybe even competing reasons. You might hold for big questions like the ones Nina brought up earlier, self-driving cars, how AI should enter into the workforce. And we like to put this word on the table because we want to differentiate it from other communication, and this is a misconception that we find students have very, very frequently, which is to distinguished arguments from fights.

So in a fight, what’s happening is it’s still communication, but it’s not relying on reasons to advocate for belief. Instead, fights use tactics. I mentioned if you hear name calling slogans, even raised voices or coercion, and I mean, if you’re an educator in the room, arguments have been kind of the bread and butter for many disciplines for a long time history, English, science, we can go on and on. And we at Thinker Analytix are trying to find ways to teach argumentation in a systematic way. And where that gets really hard and really tricky is in these kind of hot topics, topics of the moment, current events where students maybe want to make arguments to each other or are making arguments to each other, but communication breaks down.

Just to make really clear what I mean when I say argument, I mean very at a minimum, I mean a main claim, or like I said, you might call it a thesis in your discipline, a conclusion that is supported by at least one reason or what we call premises. This is the smallest possible argument. It has two claims, a main claim and a premise, and the connective tissue, the thing that holds the argument together, which is often if you’re making arguments verbally or reading it in prose, the things that are often invisible to students is that there is a connection being put forward by the author or speaker, which is that the premise is a reason to believe the claim above it. That connection we call an inference. It’s a terminology from philosophy, but an inference is part of an argument. The claims that we put forward, they’re not just disjointed kind of sentences, free floating in air when we make arguments.

We’re trying to advocate that the premise is a reason to believe the ultimate main claim or conclusion. This is just a little funny argument about whether a hot dog is a sandwich. This person is advocating that a hot dog is a sandwich on the basis of the claim that a hot dog is meat between bread. So this is just to give you a sense of the components of an argument. In particular, an argument is just claims and inferences. It’s just boxes and green lines in the way that we visualize arguments at thinker analytics. But just for now, just keep in mind claims and inferences. That’s kind of the name of the game.

Okay? There’s the basic terminology. What we do, I think analytics is try to help students clarify and get ownership of both of those components of arguments, both the claims and the inferences. The tools that we use to tackle each of those components are called systematic empathy and argument mapping. In this session, we’re going to focus on systematic empathy, and I’ll give you some pointers for more resources on argument mapping, if that’s of interest to you at the end of this. So the goal of systematic empathy, the goal of this kind of empathy process that I’ll outline in just a second is to really understand what the claims in an argument really mean, to really attend to the words that are being said and to kind of poke and question them until you understand exactly what the argument is saying, what the words are saying. We use the words systematic and empathy quite deliberately. Systematic meaning that there’s a process. So I’m going to give you a step by step process. And empathy, meaning understanding by empathy, I don’t mean agreement or even necessarily feeling in a way that maybe some SEL resources would use Empathy. All I mean by systematic empathy is just getting to just a baseline understanding of what another person is saying when they make an argument. That’s all.

Okay, so let’s talk a little bit about what systematic empathy is. As I said, it’s a step-by-step process, and this is a process that we developed and refined over time working with faculty at all different levels from grade six-ish up through college and university students and faculty. And again, the goal was to give learners a really practical, really clear step-by-step process to read or listen to an argument and get to a clear understanding of what’s being said, what the words really, really mean. So here are the steps. So when you’re encountering an argument, first, we suggest that learners pause and just identify their assumptions about the topic, even about the speaker, just to acknowledge that there are assumptions before you even jump into a conversation. We all come to arguments, especially disagreements with assumptions, and it’s important to stop and just take a pause and think about what those are.

Step two is to listen to or read the argument and try to repeat the argument in your own words. So the goal here isn’t to listen to someone’s argument and then reply instantly or object instantly and send your own ideas back. It’s just to get an understanding of what’s being said. And the way you check whether you’ve understood it is to repeat it in your own words, and then finally, ask what you don’t understand. So take a pause, ask the speaker to clarify what you don’t understand, and ask them if your restatement of what they said is right. Did I hear you correctly? Is this what you really mean?

Now, it looks like a short, short process, right? Three steps. But in practice, when students are coming to an understanding of someone else’s view, you might have to go back and forth between steps two and three multiple times. They might have many questions when they’re trying to understand an argument, and so they might need to ask, repeat, clarify, then go back and ask again, repeat, clarify, right? So this can go on for a little while, but very, the goal is just to practice these steps until a learner can restate another person’s argument and have the arguments original author or speaker say, yep, that’s exactly what I meant. You got it. Write. That’s what I mean.

So this is it. This is systematic empathy. It’s just these three steps. And what we’ve tried to do is design ways for students to practice this together collaboratively, but also independently, and to really hone these skills at their own pace. So I’m going to give you, in this session a couple samples of how you might do that with students, and I hope folks will jump in the chat and ask questions and give me your thoughts as we go along. All right? So one of the ways that we’ve tried to help students practice disagreement in the classroom and actually help faculty practice disagreement in workshops is through disagreement scripts. So often if I were to walk into a classroom and to say, okay, disagree, that doesn’t sound fun. No one wants to just, I mean, most people don’t want to just launch into a disagreement with their colleagues or their peers. So what we’ve tried to do is craft some scripts that sort of invite us to ask questions and interpret, but that are sort of on a neutral space that aren’t just students disagreeing with each other right off the bat. We’ll get there, but we start with some scripts.

So I’m just focus in on that concept of restatement, which is step two in systematic empathy. And I ask you all, so here’s a script. I’ll read it to you, and what I’m going to ask you to do is maybe in the chat, or if you want to unmute, ask how you would restate the phrase in yellow in your own words. All right. So this is a very short back and forth between two people. Person A and person B. Person A says, I can’t believe there’s so much anger around the new soda tax. More people should support it. The tax is going to help kids stay healthy and reduce our town’s plastic use. And person B says, but the new tax really hurts local businesses. We have to look out for the little guy. All right, so this is, maybe it’s controversial in your eyes, it’s maybe less controversial than some other topics I could have chosen. So I just tried to pick something a little bit of an entry point that’s policy related, but maybe not super, super hot topic. But just to ask, if you had to restate this phrase in yellow, we have to look out for the little guy. How would you do it? What would you say in your own words if you had to rewrite it?

Yeah, some of the people impacted by this might be small, local businesses, and that the impact might be negative on them. That’s why we need to look out. For sure, for sure. Thanks, Sharon. Yeah, anyone else want to take an attempt? So one thing I’ve heard folks say in these sessions before, something like they sort of read into it a little bit and they say like, oh, maybe businesses have been having an otherwise tricky time already, and so we really need to attend to policies that are going to impact them. And what’s interesting is I think for a lot of faculty, we hear the phrase for the little guy, and that’s just something we’ve been hearing for a long time. It’s sort of colloquial, it’s pretty common, but I use it because it is a little vague. It doesn’t point out, it doesn’t clarify exactly who they’re talking about. We can imply that it’s businesses, maybe we might be able to make that connection here, but that it invites, there’s at least some room for interpretation here. Great. You’re right. Yeah. So we can’t ignore the businesses run by small enterprises. Those are the ones who maybe we should pay particular attention to. So already you’re starting to see there are different ways of saying there are different interpretations you might have. And if we had 20 students in this room, we’d get 20, slightly probably different answers.

And so the point of this is not to say that there’s any right one, but that you actually can’t know the right one unless you’re in the conversation with the person and you’re able to ask. And so that’s where empathy can come in. Systematic empathy can come in. It invites you to identify these phrases that are a little bit vague, that need a little bit of clarification and actually ask what the person means. And you can imagine that I gave you, like I said, a little bit of a milk toast kind of policy topic, do tax. But if I had made this about, say, the election or a border wall or any kind of hot topic that’s happening right now, controversial topic, vagueness can cause real miscommunication because those are the moments where it’s particularly hard to attend to precise meaning. So by practicing with these sample texts, maybe topics that aren’t so, so hot, maybe even some funny topics, it’s a really great way to get students thinking about empathy and about how to practice it when the heat does get turned up. Alright.

Oops. Lemme see. Great. Here’s one more. I wanted to give you a sense of how you could also design some exercises for students around systematic empathy to practice this restatement. So here’s a new script. Social media networks like Instagram promote unhealthy standards for adolescents, boys and girls. Teenagers are deluge with photos of unrealistic lifestyles and become obsessed with appearance. Person B says, I know what you mean. Frequent social media use shows that today’s teenagers are superficial. So what’s interesting about this, the last sentence is a little vague, but it’s also potentially an overstatement. Maybe the person doesn’t necessarily mean exactly literally what they say. This happens all the time. So what we need to do is interpret the claim. Today’s teenagers are superficial in a way that’s reasonable. And so what we do in our programs is train students to interpret in a way that’s most reasonable, by which I mean is true or likely to be true or at least plausible and make sense in context. And we do that training through this sort of multiple choice exercise. So here I’ve got some options. Does it mean that A, most teenagers in the current generation are shallow. B, social media makes today’s teenagers care more about their appearance than previous generations. C, everyone born after 2000 is superficial, or today’s teenagers are thoughtful, caring, and kind. Great, I see Gabrielle posted B.

So social media makes today’s teenagers care more about their appearance than previous generations. I also had B, and we can talk a little bit about why. One thing about for A and for C, I thought the focus on most teenagers and the use of the word shallow was also a bit of an overgeneralization. I don’t know that A is true. I might question whether A is really true. Same with C, right? So in terms of reasonableness, I thought A and C were kind of questionable on a true false grounds, and D just seemed to be off the mark, right? It just seemed to be going in the wrong direction. B, what’s interesting about B is that it is more focused, right? It’s a little more nuanced. It gives some more clarity to what the person in the person B is actually saying.

Awesome. Thank you all for playing along with this. So this is just to show you how scripts can help students prepare for actual disagreements in your classroom, giving them some of these practice exercises. Analyzing scripts together can get them thinking about the actual behaviors that they need to exhibit when they are in a disagreement. And we’ve got some resources and some opportunities for faculty to see how to transition from something like this to actual classroom disagreements where students are engaging with each other and crafting their own arguments. Okay, what does this have to do with AI? I wanted to pitch a couple ways that AI can be actually really helpful in teaching these skills. I’m going to show you a couple of them here. This is by no means exhaustive. One interesting thing is that students need a lot of practice with these skills. It’s not just something that they’re going to see and just kind of intuit immediately.

It takes a lot of training. It takes a lot of practice for students to gain these skills. So here’s one example. This morning I was finishing up these slides, and after I was looking at this example of social media, I just asked, she chat GPT to make a new argument for me, one about a different topic, and they made one about video games. I asked for a student topic specific example that pokes at the same skills but uses different answers and a different prompt and it spit one out almost immediately. So I think using this as a way to craft opportunities for practice in the classroom can be really, really helpful. I’ve been doing it more and more with my own instruction and just trying to think about just different ways to craft practice, practice opportunities for students. How much time were you hoping to save for the video at the end?

Nina Bamberg (32:40):

The section of the video I wanted to show is seven-ish minutes long, but we don’t have to watch all of it. It can be kind of stopped partway through, but you get the idea pretty quickly.

Aidan Kestigian (32:55):

So if I’ve got five minutes or seven, yeah. Okay, perfect. Nice. Okay, so that’s systematic empathy. I realize that that’s a quick tort de force of just kind the steps and then how you might practice some of those steps in bite-sized chunks with your students. But that’s of course just one aspect of productive disagreement. Productive disagreements have lots of different features, and I wanted to pitch one more script to you and open up any thoughts or reflections on what makes this productive or unproductive. So here’s a slightly longer script. I’ll read it to you in case you’re on a phone. And the question is, what makes this disagree, productive, or unproductive? So here it goes. It’s about a border wall. So person A says, people have been talking about building a wall along the southern US border ever since the Mexican-American war, and it’s going to be an issue in the upcoming election.

You can’t pretend to be interested in safety, that’s supposed to say, and security, if you oppose common sense border security. Person B says what person A says, those who oppose a border wall never acknowledge that building a wall is good for everyone. It protects people already living in the United States and lightens the burdens on the already stressed legal system. Person B says, so you think that immigration is useless? Immigration helps fill dire labor shortages. Plus, have you ever read about the living conditions where immigrants are coming from? It would be cruel to send people back there. Person A says, look, immigration without proper paperwork is illegal for a reason. My uncle knows a US senator whose committee debated this policy back in 2019. Both agreed the wall is a good idea, so maybe ignore my tone and inflection, right? Maybe this was much more heated than it was originally, but just based on the text on the page, I’m curious any keywords or thoughts people have about what makes this productive or unproductive? Things that you think kind of go poorly, things that could be done better.

So there’s a couple things that I intentionally built in here. A couple things that I sort of experienced. I think teaching and curious if any of these resonate with you. There’s a couple weird things going on here. One is that there seems to be a little bit of the people aren’t really connecting on what their words aren’t kind of talking past each other, I guess is what I’m trying to say. So on the one hand, people have been talking about building a border wall for a long time. It’s going to be an issue in the upcoming election. This initial entry point at the top person A, there’s a lot going on there, and it’s not always clear exactly what that has to do with person a’s point. A lot of it is kind of background information. I’m not even sure that they really make clear what their ultimate point is. Maybe it’s that they support a border wall. Maybe it’s something more nuanced than that.

So the stress legal system seems like something productive. Yeah, that sounds like a phrase. That’s logical. Yeah. So they actually point to a reason. Ultimately, they do get to a reason for it, so that’s good. On the other hand, person B jumps from talking about the border wall to talking about immigration in general, which maybe wasn’t person a’s point. So they go from this particular issue to something a little bit more broad. So there’s a lot of stuff going on here. There’s a lot of, I think, good things, but also when I put this in front of faculty, often what I get back is a list of problems. So these people tend to focus on the unproductive.

What’s interesting is that AI can also generate scripts like this one, the one I just showed you, and the reason why that’s important. So literally what I did is ask it to remove references to the border wall and give me a new script. And the reason why that’s helpful is that scripts are a really nice way for students to practice the skills that we talked about earlier, but also to open up a conversation with students about what actually constitutes a productive disagreement. What does it look like? What does it feel like? And so you can actually ask Chat GPT that what constitutes a productive disagreement. They would probably give you a very long list of features. And I think it’s really helpful because students don’t always know what constitutes a productive disagreement, nor do they always have examples or models of what productive disagreement looks like.

Social media is not a great place to go if you’re looking for models of productivity and positivity in terms of people disagreeing but not kind of blowing up at each other. So what I’ve got here, this is just a sample, some thoughts, food for thought. This is some work we’re doing with a psychology lab at Northwestern University and trying to get it focused on some of the particular features of what makes a disagreement productive. And you’ll see, especially in these bottom two rows, a lot of what can help a disagreement be productive is making arguments really clear. That’s actually something that went wrong in the script we looked at. It was kind of unclear what the argument actually was.

And also focusing on its quality, having a shared norms with the people you’re communicating with about what constitutes a good argument. We focus a lot on arguments that have true claims and claims that are relevant, which getting back to the very beginning where we were talking about claims and inferences, we want claims that are true and inferences that are strong, something we take from philosophy. So this is not to say that these should be your norms for your learners, nor should it be an exhaustive list by any means. But just wanted to put in front of you some of the kind of characteristics, some of the norms that you might explore with students and the ways that you can actually craft scripts to exhibit or violate these norms. And actually, if you go back and check, the border wall script does actually violate each of these to some extent.

And you can actually find the places where it does that. Just to close, I’ll just say, and this is in the deck too, border wall might be too hot a topic for your learners. This is exactly the same structure of an argument, the same, sorry, structure of a disagreement. The shoes of the doors slide. It’s actually exactly the same back and forth. It’s just with a different topic about whether you should take your shoes off when you go inside your house or not. This one was actually inspired by some college students making disagreement samples for us. So it’s supposed to be somewhat comical, somewhat outlandish. So what you can do is actually write scripts and use chat GPT to help you that where you change the kind of heat, you can vary the heat of the conversation by changing the topic, but by still targeting some of these norms, basically keeping the structure the same.

Alright, finally, I’ll just stop by just saying we’ve got a bunch of stuff coming up that may be of interest to folks on systematic empathy. There are links here for folks in K 12. I know we have some K 12 folks in the room. We’ve got a course running through Harvard School of Education this spring. And we also do in-house workshops where we get a much deeper dive on both systematic empathy and argument mapping and all the links are here. And finally, our resources for students are here as well. Think arguments is like our signature program for students. It’s online. It teaches them systematic empathy and argument mapping through asynchronous training. And basically they can take it at their own pace. They can do it for homework outside of class, just trying to reach learners everywhere as robustly as possible through a mastery learning platform, which the folks at Pedagog.ai have had a huge hand in helping us make and deliver and think puzzles, which is a game with quick fire practice with arguments. Yeah. So I just want to put those here and folks can kind of have a look through these as we finish up and I’m happy to answer questions at the end as well.

Nina Bamberg (43:19):

Yeah, thank you. So we just wanted to wrap up with a quick, kind of back to combining everything we just talked about. So the important arguments for the age of AI as well as the systematic empathy scale. And I was hoping that we could watch a, it’s just a few minutes of a real hearing that took place in Congress, in the Senate, and maybe we can watch this argument happen and discuss what was either productive or unproductive in this conversation, in this back and forth based on what we just learned about systematic empathy. And then we could, I don’t know, Aidan, if you had ideas on incorporating this into something like this, into teaching.

Video (44:27):

I’m the only one sitting here, it’s bad news for the bad news for the witnesses change. I’m sure. Lemme, Mr. Lemme come back to this. We were talking about kids and kids’ privacy and safety. Thanks for the information you’re going to get me. Lemme give you an opportunity though, to maybe make a little news today in the best possible way. 13, the age limit for Bing chat. That’s such a young age. I mean, listen, I’ve got three kids at home, 10, 8, 2 are my kids. I don’t want my kids to be interacting with chatbots anytime soon at all. But 13 is so incredibly young. Would you commit to today to raising that age? And would you commit to a verifiable age verification procedure such that parents can know, they can have some sense of confidence that their 12-year-old is not just saying to being, yeah, yeah, yeah.

I’m 13. Yeah, I’m 15. Sure. Go right on ahead that now let’s get into it back and forth with this robot. As Senator Kennedy said, would you commit to those things on behalf of child safety today? Look, as you can imagine, the teams that work at Microsoft, let me go out and speak, but they probably have one principle they want me to remember. Don’t go out and make news without talking to them first, but you’re the boss. Yeah. Let’s just say wisdom is important and most mistakes you make when you make them by yourself, I’m happy to go back and talk more about what the right age should be. You think 13 is awfully low though. It depends for what actually to interact with a robot who could be telling you to do any number of things. Don’t you think that’s awfully young? Not necessarily.

Let me describe really is the scenario. When I was in Seoul, Korea a couple of months ago, we met with the deputy prime Minister, who’s also the Minister of Education, and they’re trying to create for three topics that are very objective, math, coding and learning English, a digital textbook with an AI tutor so that if you’re doing math and you don’t understand a concept, you can ask the AI tutor to help you solve the problem. And by the way, I think it’s useful not only for the kids, I think it’s useful for the parents and I think it’s good. Let’s just say a 14-year-old, let’s say, what’s the age of eighth grade algebra? Most parents, I found when my kids were in eighth grade algebra, I tried to help them with their homework. They didn’t believe I ever made it through the class. I think we want kids in a controlled way with safeguards to use something that way. We’re not talking here about tutors. I’m talking about your AI chat being Chad. I mean, famously earlier this year you were chatbot, you had a technology writer for the New York Times who wrote about this. I’m looking at the article right now. Your chatbot was urging this person to break up his marriage. I’m not sure we want 13 year olds to be having those conversations. No, of course not. Okay. Which is, well, will you commit to raising the I actually don’t want bing, Chad to break up. I don’t either, but that might be some exceptions.

Yeah, but we’re not going to make the decision on the exception. No, but it goes to, we have multiple tools. Age is a very red line. It is a very red line. That’s why I like it. And my point is, there is a safety architecture that we can apply to bring a, but your safety architecture didn’t stop. An adult didn’t stop the chatbot from having this discussion with an adult in which it said, you don’t really love your wife. Your wife isn’t good for you. She doesn’t really love you. Now this is an adult. Can you imagine the kind of things that your chatbot would say to a 13-year-old? I mean, I’m serious about this. Do you really think this is a good idea? Yeah, but look, wait, wait a second. Let’s put that in context. At a point where the technology had been rolled out for only 20,000 people, a journalist for the New York Times spent two hours on the evening of Valentine’s Day ignoring his wife and interacting with the computer, trying to break the system, which he managed to do.

We didn’t envision that use and the next day we had fixed it. Are you telling me that you’ve envisioned all the questions that 13 year olds might ask and that I as a parent should be absolutely fine with that? Are you telling me that I should trust you in the same way that the New York Times writer did? What I am saying is I think as we go forward, we have an increasing capability to learn from the experience of real people and put the right, that’s what worries me. That’s exactly what worries me is what you’re saying is we have to have some failures. I don’t want 13 year olds to be your Guinea pig. I don’t want 14 year olds to be your Guinea pig. I don’t want any kids to be your Guinea pig. I don’t want you to learn from their failures. You want to learn from the failures of your scientists.

Go right ahead. Let’s not learn from the failures of America’s kids. This is what happened with social media. We had social media who made billions of dollars giving us a mental health crisis in this country. They got rich, the kids got depressed, committed suicide. Why would we want to run that experiment again with ai? Why not raise the age you can do it. First of all, we shouldn’t want anybody to be a Guinea pig. I think regardless of age or anything. Good. Well, let’s roll kids out right here, right today, right now, no, but let’s also recognize that technology does require real users. What’s different about this technology and which is so fundamentally different in my view from the social media experience is that we not only have the capacity, but we have the will and we are applying that will to fix things in hours and days.

Well, yeah. To fix things after there’s been, after the fact. I mean, I am sorry. It just sounds to me like you’re boiling down. You’re saying trust us, we’re going to do well with this. I’m just asking you why we should trust you with our children. I’m not asking for trust, although I hope we will work every day to earn it. That’s why you have a licensing obligation. There isn’t a licensing obligation. That’s why in your framework, in my opinion. Well, sure. But I’m asking you as the president of this company to make a commitment now for child safety and protection to say, you know what? Microsoft is going to, you could tell every parent in America now, Microsoft is going to protect your kids. We will never use your kids as a science experiment ever. Never. And therefore, we’re not going to target your kids and we’re not going to allow your kids to be used by our chat bots as a source of information if they’re younger than 18.

But I think you’re talking about, with all due respect, there’s two things that you’re talking about and I think we’re, I’m just talking about protecting kids. It’s very simple. Yeah, no, but we don’t want to use kids as a source of information and monetize it, et cetera. But I’m equally of the view. I don’t want to cut off an eighth grader today with the right or ability to use this tool that will help them learn algebra or math in a way that they couldn’t a year ago. Yeah. Well, with all due respect, it wasn’t algebra or math that your chatbot was recommending or talking about when it was.

Nina Bamberg (51:47):

Okay, that’s about where I wanted to get to in the video. But based on what we talked about today, where do we see some productive or unproductive argumentation skills exhibited here? Or even for the sake of time, I don’t know, Aidan, if there’s anything that stuck out to you.

Aidan Kestigian (52:18):

No, it is interesting because to think about what each person’s kind of goal was in that conversation, one person is clearly defending a corporation or their kind of track record or trying to explain I think a little bit what’s going on. Yeah, there is some argument and counter argument. Certainly. I do wonder to what extent clarity was the goal if both folks were actually interested in understanding the full nuanced view of both sides. I think some of the stuff about testimony is that it’s very short, it’s very fast. It’s not necessarily designed or publicized in a way that makes it fully transferred knowledge, I guess I would say. And so one thing you could do with students I think with this video is sort of ask the same questions, but then ask if you had to restate the argument that’s being made on either side, what is it? Has there been enough clarity in what questions would you ask if you needed more clarity? I think that would be super interesting. Yeah. So that’s what I would do.

Nina Bamberg (53:53):

Yeah, no, I totally agree. Right, because this is a great example, Gabriela, right? It seems like they’re possibly defending different points as Aidan kind of said, both sides seem to have different goals here in that conversation. But watching something like this as a real life example could be a great way to take once you’ve learned these skills of systematic empathy and taking it to the next step and recognizing that it’s not always going to be a neatly laid out back and forth script conversation in the real world, and it’s going to be harder to parse through arguments that we encounter out in real life. And this is a great example of something like that. One thing, I know we’re at two o’clock, but one thing I just wanted to quickly end with is a couple more examples on using AI to build some of the skills that we talked about today.

So the first thing I’m going to show is an example of our debate, or sorry, our chatbot that we have at pedagog.ai called socrat.ai if you want to check it out. And one of the tools that we have on socrat is a debate tool. And so you as the teacher could set the debate topic and your students could go in and have a debate with Sorat on that topic. And what’s great about the bot is that it’s programmed to do two really important things. One model good argumentation skills. So he always responds kindly saying things like, I understand your point, and then laying out his argument in effective ways. And he always sticks firmly to the other side of what the student is saying in order to challenge the student on their own thinking. So that’s just one example of how you might use an AI tool to teach argumentation.

And another couple things I wanted to show is examples just from regular chatbots like chat GPT on how the chatbots could be used to help students explore multiple perspectives. So this is an idea here where I acted as a student and said, I’m participating in a debate about banning junk food in the school cafeteria. I’m arguing in favor of the ban. What might my opponent argue And chat GPT gave me six good arguments here for something that I should consider if I’m engaging in this debate for understanding what my opponents might say. And it’s a great way for students to make sure that they’re looking at multiple perspectives of an issue because we see obviously, that they can often get stuck in their own way of thinking or in their own argument. And the AI tools can be prompted to help them explore other ways of thinking about an issue that they care about. And here’s another interesting example where on the teacher side of things, I said, help me plan a debate on wind power by suggesting sources on both sides. For this, we used perplexity.ai, I believe because it cites its sources very helpfully. And so here’s another example of how you might do that, what you’re looking for and you want some sources to help you back it up again, help make some of those arguments on both sides of an issue if it’s getting too stuck in one particular realm. Okay.

Zahi, I’m curious what you mean by that question. Do you want to share a little bit more, or what do you mean by limiting their potentials?

Speaker 5 (57:52):

Yeah. Hello? Can you hear me guys?

Nina Bamberg (57:55):


Speaker 5 (57:56):

Yeah. Yes. I was thinking that when students have use this prompts like these chat GPT or any other AI tools, I think that as we are saw just bank, the debate between the senators at what age they should be allowed to use prompt or this technology because think of our time when you’re going to high school. We had no such tools. We had to think among ourselves. We had to talk between ourselves, socializing with our friends, even with our teachers about the debates about the topics. Now know we simply giving them a kind of a tool. You type the prompt, you get the responses and you choose which one you think is better, and then you about deliver your talk or your writing. So in what way? You tell me you can explain. Do you have any kind of evidence that okay, this way of learning is actually helping students already kind of potential they can expand their creativity or their thinking capacity.

I mean that’s so important because I teach online and we are always concerned how students are writing online. Are they using any kind tools behind the camera? I mean, we don’t know what they’re doing behind the camera because I can see their face, they’re very good looking good, but what they’re doing with the keyboards, are they using another window and using that in their writing production that be sometimes difficult to check because still there is no reliable software or any kind of tool that can detect that, okay, this plagiarized or copied or produced by chat GPT resources. That’s my take on that. Yeah, that’s my concern

Nina Bamberg (59:50):

Priten I saw you turn your camera on. Did you have something you wanted to jump in and say? Yeah,

Priten Shah (59:57):

I mean, I think I was just going to start to think about, maybe we can think a little bit out loud about the fact that I think there’s some real concerns about the ways that we might limit students’ thinking as we see students to use the tools in ways that are not productive. But I think that’s always been true at all different levels of technological releases. So when the calculator initially came out, and I know that’s a go-to example for everybody, but I think it’s a great way to just think about this. We were really worried about students’ mathematical skills and the kinds of intuitive mathematical skills we’d build in students if they were able to rely on calculators at early ages. And so we’ve scaffolded when we introduced calculators within our school systems. So we don’t necessarily, by fifth grade, you use only maybe arithmetic calculators.

And by 12th grade and calculus classes, you’re using a much more graphic calculator. And there might be equivalent ways we approach AI integration within our school systems as well, where maybe we start with our kindergarten students and they’re not using any AI because we want them to develop some of those basic skills and kind of start to learn maybe even empathy skills or augmentation skills. And then maybe by fifth grade they’re using some basic chat bots that kind of are guardrail and they can have particular debates on topics their teachers have instructed them on. And maybe by 12th grade they’re now using full on public like the adult tools and they’re able to use it for research and they’re able to use it for outline their argument. They’re able to use it to think about the cognitive biases they might be inhibiting, inhabiting in their own thinking and writing.

And so I think when we think about these tools, I think they’re dangers that come from students using them in the wild. And there’s ways that we can actually integrate them to actually help students think further and use them in productive ways. And I think our jobs as educators and start thinking about how can we use these tools to better help them argue better, help them with whatever skillset we want them to develop through our classes, but also recognize the reality that they will use these tools, right? There is no limiting their access no matter what the senator pointed out, a very good argument that, or just a very good fact that right now there’s not even the age limits on all these platforms. It’s kind a parce because you can kind of go in and sign up by just saying, you’re over 13 and whatever new barriers they put in place this, we’ll find ways to get around them.

And so I think having these conversations with our students about what productive uses are, helping them navigate some of these questions and helping them think through the reasoning skills themselves to kind of think through what are productive uses for me? What are ethical uses that I want to use AI in right now? Might I want to use AI in the future in my future career paths? And what impact does that have on society? Those are the kinds of questions we want to start. And these are deep questions. They’re not easy questions, and this is not an easy task for us to do as educators, but I think this is the reality of what lies ahead of us. I don’t know if there’s anything else you wanted to say on that.

Nina Bamberg (01:02:36):

Yeah, no, I just wanted to make the point on something like this in terms of one of the points you made that we used to, or the expectation was that students would learn different perspectives from their teachers or through in-person discussions and things like that. We, Priten and I have taught many, many students on writing and making arguments and counter arguments, and we find almost every time that the counter argument tends to be where they get the most stuck. They know what they want to argue, they know what they think and what their opinion is and what their main claims are. So think trying to figure out what the opposite that would be, what a reasonable person who disagrees with them might say can be really hard. It’s very easy for us as humans to assume that if someone thinks differently on an issue, that their beliefs must be the most outlandish option possible in order to be seeing the issue differently.

And so using the tools in ways like this by asking it for what an opponent might argue is a great way to show, as Aidan would say, true reasonable claims that are opposite of what the student is currently thinking. And so I guess in terms of assuming that in-person interactions are always better for getting at all sides of an issue, I would say that I disagree that that’s true because there are so many nuanced views out there and it can be really hard for students and for anybody a lot of the time to write, paint their opponent’s view in a reasonable light. And AI is great at that.

Priten Shah (01:04:37):

Yeah, I think the goal thing about how can we use AI to better interact with other humans rather than replace our interaction with other humans, and this is a great example of that.

Nina Bamberg (01:04:51):

All right. I know we’re way over time, but any final questions, we can definitely stick around for a couple minutes if you have any final questions. But Aiden, I can share your slide decks, right? Oh yeah, for sure. Feel free to. Okay, great. So both slide decks as well as the recording from this webinar will be posted soon in the same place that you accessed the Zoom today, so you can always come back and have access to the resources. Thank you all so much for coming. Thank you everybody.

Sign Up For Our AI & Education Newsletter!