Can AI tutors fix what schools can’t? – with Vida C. Williams

Kevin P. Chavous: what makes a great tutor? Patience, consistency. Someone who sees how you think and helps you think better. Now, imagine that tutor isn’t a person.

It is an algorithm in districts across the us. Artificial intelligence is being tested as a personal tutor, which includes grading papers, customizing lessons, even flagging  when a student starts to fall behind. Vida Williams is Chief Data Officer at Stride, a national education company working to bring AI tutoring.

Into everyday classrooms. She’s helping to shape the design and the guardrails of what could become a new standard in American learning. So what happens when teaching becomes a shared task between humans and machines? This is what I want to know.

Vida Williams, I tell you, I’m excited to have you on the show. You’ve had a remarkable career. Um, you are a data science genius. Now that’s a language I don’t quite understand, but I know you. You’ve got game in that space. So that’s a good thing. Uh, we wanna talk about the AI tutoring concept and what it means for  American school aged children.

But first, let’s talk about you. Uh, I know you’ve had, uh, an illustrious career in healthcare and energy, but why education after all these years and all you’ve accomplished, why did you decide to devote this part of your career? In the education space?

Vida Williams: Uh, I think it’s two reasons. Um, the first reason being that I’m to kind of in the twilight of my career, I don’t think I’m quite done yet.

I’ve many, many good years left on this dread, I would say. Yeah. Um, but. Knowing that you’re in a twilight means that you’re wanting to make sure that you’ve left something behind that’s gonna make the world better, right? Especially if you’re an engineer, especially if you’re in technology. You wanna make sure that we are passing a, a good framework.

Forward for solutioning. And I could think of no better place to do that than in education. Right? Um, education  provides a legacy opportunity for me, and that’s the most selfish response I can give you, is there’s no better place I wanna be associated with than having made a difference in education. I think more tactically, however, why education.

Education and technology right now are both in a transformative state. Both industries right now are undergoing an extreme amount of change, and it is a reflection of where we’re gonna be tomorrow on the care that we give to the kids today. And I wanna be part of making sure that when we’re utilizing technology.

In their lives to the intimacy of education and the way that that’s gonna impact them, that we’re doing so responsibly, that we’re doing so ethically, that we’re doing so thoughtfully. And I trust myself to do that. I don’t know if I trust everybody. Does. Does that make sense? Yeah. Well,

Kevin P. Chavous: it makes a lot of sense.

And, and,  and this, this notion of your timing in terms of moving into the education space is really important because I don’t know that people fully appreciate the level of transformation and change that will exist. In the everyday learning space. I don’t even talk about education. I’m talking about learning as a lifelong advocation for people of all ages.

Uh, and you know, it’s, people sort of got a whiff of it with chat GPT, but now this whole artificial intelligence and AI. Teaching tools and the like. Uh, they’re gonna come forward even in a more prominent way, uh, over the next several years. Uh, you are the chief data officer at Stride. What do you do in that role?

And then, then we’ll talk more and more about, uh, the AI tutor experience.

Vida Williams: So I think more and more companies are moving. Towards a, a chief data officer, there is a lot that we  had yet to consider.

Kevin P. Chavous: Um, let me jump in for a quick moment. For a lot of students, the traditional model just isn’t the right fit.

That’s why more than 3 million families have chosen K 12 power schools, offering flexible tuition free learning options with certified teachers, and a personalized approach. That meets learners where they are. If that sounds worth exploring, you can head to k12.com/podcast. Alright, back to the conversation.

Vida Williams: When we started going full tilt into InfoSec and those types of positions, it took a minute for us to realize that the reason why information security was so important is because the data was so sensitive. And so a lot of what I do at Stride is protecting our children’s data. So it’s organizing it in such a way that we can have clarity on it.

It is making sure that we are implementing good governance  practices, that we are. Following ferpa, kapa, HIPAA compliances within this organization. So those are aspects of my job that are super important that maybe people don’t consider a lot because what they like to think about data. Is the reporting or how we’re using it, and now more increasingly how it’s being used to generate AI capabilities.

But there is a very solid aspect that I think more and more companies are acknowledging a need for, and that is ensuring our, our safety and security of, of information.

Kevin P. Chavous: Well, let’s talk about the, uh, use of artificial intelligence. As I said earlier, uh. During sort of the advent of chat GPT? Yeah. Uh, more and more folks started to sort of put their toe in the water to see what it was like.

And with teachers, for instance, uh, the use of ai, uh, has been pronounced  because, uh, a lot of folks who I know that are teachers have said that their biggest challenge. Was the administrative responsibilities, the duties, the tasks, the reporting, and one thing that AI has done has made all that a lot more manageable, especially once teachers get comfortable with the tool.

But now that everyday people are using it, not just in the teaching profession, but in other jobs and students are using it, when did you start to think in your role? That this could really be a benefit for students in terms of aiding them on their learning journey?

Vida Williams: Honestly, because I grew up in data, Kevin, I thought about this, I wanna say a decade, maybe a decade and a half ago.

Kevin P. Chavous: Yeah. So in other words, you were well ahead of. All the rest of us. I had to be, though I wouldn’t be good at what I do. Um, but lemme

Vida Williams: tell you how  I was thinking about it, because I, I think it’s one of the things that is going to be special, right? There’s a, there’s a lot of things about the, the movement of AI that makes us nervous, right?

Like, especially teachers and students. And you think about, you know, the epidemic of, of cheating or, hey, they’re just gonna use. Chat GPT to write their papers. Like there’s a lot that makes us nervous, but one of the things that makes it almost special is they re they, I’ve called it the resurgence of library sciences in its truest form.

Teachers will have an opportunity, educators will have an opportunity. Lifelong learners will have an opportunity to become curatorial instead of just creators, right? So they’re ab absolutely going to have an opportunity to really have a curious moment and deep dive in it, and see the wealth of information that’s at their  fingertips, and then begin to complete.

Pile that information into a mosaic, if you will, of deep learning. That is a part of AI’s capabilities that we aren’t quite highlighting enough because we’re only looking at the automation capabilities and, and let’s put a pin in that ’cause that is also special. But this curatorial capability is gonna be a new framework of.

Needs that is rooted back in the library sciences. And as a girl whose grandmother was a librarian, I can see her coming full circle in this story. Right? Yeah. Um, and I just, and I just love that aspect of it. So,

Kevin P. Chavous: so let me ask you this, Vida, because that, that’s the, that’s the fascinating perspective. Uh, and to break it down.

Um, so in other words. Uh, if, if a, if a student or a teacher or anyone who’s learning, but let’s say a student has a thought in real  time, then AI allows them to, to flush that out, if you will. Uh, or if they are stuck, you know, in the middle of a lesson plan or, you know, or something like that, they can flush it out.

And correct me if I’m wrong, what you’re saying is not only can they. Sort of, uh, address their challenge, but it also could, could help feed the curiosity that comes as you’re in the midst of your learning journey now. That’s right. I mean, I didn’t say it quite as eloquent as you did, but I, I get the sense that you’re saying that this is real time usage that could help a kid not only.

Fix what ELs them in terms of their understanding, but also take it to the next level.

Vida Williams: Absolutely, and I would say more specifically when we’re talking about personalized learning, let’s be honest, we don’t all like all the subjects that we have to learn in school, right? Like some people just really did not like math.

Like I wasn’t one of those, obviously. I kind of liked that. Right,  right. But if I didn’t, then a personalized learning journey along these same lines, a curated experience means that I get an opportunity to learn that more efficiently and more expedited so that I can spend more time learning things that I’m just insanely curious about.

I wanna know what it was to be, you know, in. Senegal in the 17 hundreds and then deep dive where the mythology of, you know, Scandinavia, right? Like whatever it is that I might be more interested in those, those pathways of learning open up. But not at the detriment of the other stuff that maybe I’m not as interested in because that can be made more efficient knowing Exactly.

I need to learn more about fractions, so it’s gonna teach me fractions today, but I don’t have to spend hours on fractions ’cause it’s figured out  how I learned. And that this is the best way to get me to learn this. Does that, does that make sense?

Kevin P. Chavous: It makes sense. And let me ask you how practically the, the tutoring aspect of this will work because right now, uh.

It’s unfortunate, but the data hasn’t changed in the last 20, 25 years. Two thirds of American high school graduates are, uh, not grade ready when they graduate. In other words, they don’t have the competencies in math, science are reading, and so they’re behind. Yeah. Two thirds of American high school graduates.

That hasn’t changed. Uh, they’re not at grade level. Uh, and many students, because, you know, the average student on their learning journey needs additional help beyond classroom. Uh, you know, parents, you know, utilize tutors mm-hmm. To help try to fill in the gaps. Uh. School systems provide  them parents reach into their pocket and pay for them.

There will be parents who are listening to this show across the country. Talk to me about how practically AI would work right now. If a parent needs a tutor mm-hmm. Then. They, they try to go through their school system. There are companies out there that provide private tutoring, but there’s a process.

Yeah. But if you have, and, and sometimes you have to wait. Yeah. Or, or go through different, you know, experiences to find the right match. How would this change with AI and practically, how would an AI tutor be integrated into a child’s sort of. Learning experience.

Vida Williams: I think there’s two ways. Um, and so I’m gonna use a couple terms to differentiate them.

Sure. Uh, I say on one side there is what we’re hearing, the buzzword around agentic capabilities. Right. And what that is, is automating a task well. When you’re thinking about a tutor and you’re thinking about  the ability to create exercises around a particular subject, then that is really automating kind of a workflow of, of just the practice of it.

So that would really. You know, if we’re talking about math problems, I remember when we were kids, we had the hundred to do a hundred math problems in five minutes and we used to get stars if we did all, no, I, I

Kevin P. Chavous: didn’t have, I didn’t have that ability or that I, I, I, I don’t, I, I don’t know. I, that was a trigger for me when you said that,

Vida Williams: but you remember when you said,

Kevin P. Chavous: yeah, that was a trigger sheet I used to.

And all of that. Yeah, I do remember. Well, but I, I’m sorry to interrupt, but it’s just you. No, no, no. I, I needed a moment and you did, because I was an

Vida Williams: overachiever, so I would always try to do all four disciplines on the first day of school so that I didn’t have to think about it anymore. Kevin, I listen.

I’m a weird person that way.  So when we’re thinking about the use case that way from an exercise, from just a practice perspective, AI can do that and, and do it in a personalized way, meaning, okay, so she’s got this. Level of maturity in this competency. Let’s see if we can press her a little bit more. And we’ve already seen some of that with the way that they’re doing adaptive testing now, right?

Like just going as far as you can and it, so it’s, it’s a step up from that, from an automated perspective. The other part of this though, is what I would consider conversational companions. Sometimes the, the method of learning is more along the pedagogical lines of Socratic method, right? Are we asking the right questions?

Do we know how to write, ask the right questions? Mm-hmm. Are we thinking critically about the answers we’re providing? Is there more critical thinking that needs to happen in that space? And thus asking the student to unpack how they’re  thinking about something. That companion, that conversational companion piece Yeah.

Is a different aspect of ai and when you put both of these things together, you now have a fullness of experience and learning capabilities. I’m sure there’s, you know, finer points in there I can make. But that is kind of where we’re gonna see the, the largest soonest innovation in the use cases of AI and tutor.

Kevin P. Chavous: You know, as I hear you talk, I, I’m, I’m reminded of the fact that, um, for, you know, the OGs out there, the older generation, uh, it, some of this may sound a little foreign, but uh, for a lot of young people who are in the midst of their learning journey, they’re so conversant and used to technology Yeah. In ways that many of us.

Don’t understand. Um, and studies have shown that in the 12, 13, 14 age group, uh, you  know, the AI companion either as part of the learning journey or just as a friend, if you will, is these young people are more receptive to that. And, and, and I raise this because there’s this issue of teachers. Being viewed as trusted voices or trusted ears for students.

Yeah. And parents have generally been comfortable with that, but we’re also, if you will, uh, vital to, we’re moving to a place where, because of the technology, AI can be a trusted voice for today’s student.

Vida Williams: Let’s unpack that a little bit because I, I think. There’s, there’s measures we have to remember. AI is engineered.

Okay.

Kevin P. Chavous: Yeah.

Vida Williams: Because it is engineered, we can put, uh, programmatic guardrails in place. So let’s say that this, yeah. Utility should only be talking  about history and, and or we’re a Spanish tutor. So we’re just gonna talk about these things at this level. Um,

Kevin P. Chavous: yeah,

Vida Williams: what I think parents need to be aware of is that it.

Is controllable. It is something that they need to check on when they’re finding utilities that they might be interested in trying for their student to really investigate the practices that were used to build the utilities. I know on my teams we have a do no harm edict. As a data scientist, we have a do no harm edict, and so therefore when we are building these, we build these utilities.

With the child’s experience in mind. That is an evolutionary step that came out of some of the controversy that we had even with Instagram, right? Yeah. When we really started with social media, we couldn’t anticipate how students or young people were gonna respond to  that type of technology. We now know.

How they responded to that, and it gives us some indication on what we can expect and what we need to guard against as we continue to, to evolve this capability. D, does that make sense?

Kevin P. Chavous: That’s an excellent point. In fact, I’m glad you used the word guardrails because that’s where I was headed, that what?

What are the safeguards that can be put in place to make parents feel more comfortable, even if the kids themselves? Are more inclined to be accepting of, of, of the AI tutor experience. And you answered it in, in, in a full way, which I appreciate. And teachers, as I understand it, they also have to be trained to make sure that they are involved in the process and you’re not just turning it all over to the algorithm.

Vida Williams: That’s right. I mean, I, I don’t think. I, I will be honest, I am not  of the opinion that AI is going to take over for human interactions. I, I don’t sit on that philosophical line, Kevin. There are peers of mine and counterparts that do. And I think it requires all of us to have these philosophical differences in order for us to have the friction to innovate fully.

What I will say is because I think that the human interaction is one of the most special. I often say dolphins teach dolphins, right? Like you have to be of a kind to teach a kind. Um, AI capabilities are engineered capabilities. We’re developing them and teachers have to be part of the process of engineering, or we are not going to get the benefit of the years of experience that they’ve brought to understanding how a child learns and engineering with that in mind.

So when we are  doing it on my teams, we work with teachers. We don’t work outside of the teacher element. We work Yeah. With educators to explain to us why certain aspects are important. What would happen in some cases if we omitted a certain interactive style and what interaction styles cannot be emulated, right?

Kevin P. Chavous: Yeah.

Vida Williams: Um, so we work alongside the industry so that we are fashioning a utility and we have to remember it’s a utility. We’re fashioning it in the likeness of. And then hopefully extending the competency forward with an innovative capability. I. That’s, yeah, that’s,

Kevin P. Chavous: that makes a whole lot of sense. And, but when you speak about utilities, I, I just have a, a couple more questions about it though.

I could talk to you all day about this stuff really. Uh, but when you talk  about utility, a tool, a valuable tool, we’ve had a history in this country of not making sure that all have equal access That’s right to those tools. That could help kids learn. That’s right. So when you, when you talk about access, and I’m talking about, you know, diverse learners, rural population where there’s huge issues in terms of gaining access to resources that will help them in their learning journey, what is going to be put in place to make sure that there’s equal access to the tools like AI Tutor that will work for kids.

Vida Williams: I am gonna make it a little complicated before I make it a little simple, because one of the big hurdles that we have for AI is that it doesn’t necessarily understand all of our dialects. I. Yeah. So if we think about our Appalachian kids, if we think about our Creole kids, like it’s not just children in a foreign language, I  think we’re getting the foreign language part.

You know, whatever disposition of foreign language is, um, what we’re missing are dialectic differences. So when our students are from West Virginia and speak a dialect that hasn’t been incorporated into ai, then they have a hurdle. We have to do better at making sure that there is a wider range of di dialectical, um, sensitivity, especially in a country as diverse as the United States that has so many pigeon languages because we are from all over.

And, and then we are also, um, geographically located. So that’s one side of it that I know we are. Actively looking at how we can better engineer for that. Um, the other part of it though, I think is part of a broader initiative that we have to have as a community  of lifelong learners to make sure that all of these utilities are democratically available.

Kevin P. Chavous: Yes.

Vida Williams: Does that make sense? Like it has to be a right, that these things are made available for the people.

Kevin P. Chavous: Yeah. And I think you’re absolutely right and I also think that, uh, I know the Stride works with over a thousand school districts. That’s right. Those that are, uh, working on this tool and working on making it more practical and usable for today’s students.

Uh, and it will be involved in. Working with school districts and training school districts, it’s incumbent upon all the folks who are, are on the front end of this, that they engage with people on the back end. So it’s universally used. That’s right. Uh, vital. I have one more question. Mm-hmm. It’s, it’s, it’s more of a philosophical question.

This is what I really want to know. Where is all this headed? I mean, project for me this the future  marriage of humans and machines. Yeah. When it comes to. The learning journey of America’s children.

Vida Williams: It’s funny that you should, that you should ask that question. ’cause I’m, I’m working through a theory. I honestly think there’s gonna be several different pathways of, of innovation trajectories.

I don’t think it’s just gonna be one. And I, you know, when we were coming up, and I’m gonna say this again because, you know, we both have gray hair. Um, you know, we had the vision of the Jetsons and everybody was living up in the skies and you had to wonder what was going on with all the people on the ground.

Like who, who stayed on the planet, right? And one of the, the, the thesis there is that everybody was living in a, like a dystopian on the ground. I’m like, no, you have whole swaths of people who are, are finding innovation back back. In what I consider indigenous to the planet, right? Where they’re looking at agriculture, they’re looking at global warming,  they’re looking at sustainability and they’re creating ecosystems.

There’s going to be a technology deployment to enrich the planet as a solution. As much everywhere as it is to, you know, enable us to go into the sky as much as it is to enable us to go into the ocean. Technology, I think, is going to become. Yes, more we’re gonna become more, uh, dare I say Cyborg, but all of us are walking around with phones.

Taped to our hands right now, and I don’t know anybody’s phone number anymore. And so you know, that idea that we’re allowing machines to aid us in the quality of our living experience and to solutions for some of the challenges that we’re heading into, I think is absolutely the best case scenario. But going back to those guardrails, we better be intentional.

About what we’re putting out there and we need to all consider what the  best solutions are, not the few people for all of us that does that make sense?

Kevin P. Chavous: It makes a lot of sense. Well said. Vi Williams, thank you so much for your insights and the work you’re doing on behalf of our children, and thank you for joining us on what I want to know.

Vida Williams: It was absolutely a pleasure, Kevin. I hope you have me back.

Kevin P. Chavous: Thanks for listening to What I Want to know. Be sure to follow and subscribe to the show on Apple Podcast, Spotify, or your favorite podcast app so you can explore other episodes. And dive into our discussions on the future of education and write a review of the show.

I also encourage you to join the conversation and let me know what you want to know using hashtag WI WT K on social media. That’s hashtag WI WT K. For more information on Stride and online education, visit stride learning.com. I’m your host, Kevin p Chavis. Thank you for joining.  What I want to know.