123: Voices Unveiled: AI and Education, A chat with Student Voice's Founder Stuart Grey

Holly Owens (00:02):

Hello everyone and welcome to another amazing episode of Ed Up Ed Tech. My name is Holly Owens

DaNadia Johnson (00:11):

And my name is Nadia Johnson and we're your hosts

Holly Owens (00:15):

And we're super pumped. We have an awesome guest with us today who is based out Scotland. We have Stuart Gray, who is the founder of Student Voice with us. Stuart, welcome to the show.

Stuart Grey (00:27):

Oh, hi. Brilliant to be here. Thank you for the invite.

Holly Owens (00:30):

Well, we're excited to chat with you and get to know everything about Student Voice so that you can share it with our audience. But before we do that, we want to know a little bit about you, Stuart. Tell us your background. How'd you get into this industry, become a founder of an EdTech company, give us all the details.

Stuart Grey (00:46):

Yeah, no problem at all. So everyone's got their own unique EdTech story. Everyone's got their own winding path. Mine came from academia. I was always a sci-fi geek as a kid and loves space and planes and things. So I ended up studying aerospace Engineering. So you're like Star Trek and things like that. Full On. Full on, full on, full on, yeah. Unabashed. And so aerospace engineering seemed the obvious thing to me as an 18-year-old. Decisions are made. And after that I worked in engineering for a bit, but then my head was turned by doing a PhD in aerospace engineering, particularly in the space stuff. So I got a chance to work on some orbital mechanics research as part of my PhD, which I absolutely loved. I did that help here in Glasgow. And then once I'd got my PhD, moved down to London to do my postdoctoral work, which was on analyzing the orbits of satellites that measure sea level rise for global warming, sort of monitoring.

(01:50):

Super interesting tons of data, very important and absolutely loved it. Then I got my first permanent academic position where teaching becomes a thing. The little ivory tower suddenly expanded a little bit and it turns out I really enjoyed doing it. I absolutely love teaching. I really enjoyed teaching what I know about engineering, mathematics, those sort of things to students. And what I brought to it was these approaches from engineering, my experience of problem solving and using data to actually drive decisions, things like that. A bit of rigor in that sense. And when I started teaching, I would teach 350 students, undergraduates, and I'd have no idea if I was good or bad or indifferent, and I thought I need some data on this. And that sort of started that journey. The early thoughts of we need to do something around better use and capturing of data around teaching in order to support educators doing what they do.

Holly Owens (02:53):

People don't realize how much data impacts the process. And a lot of, I also teach and Nadia's been a teacher in the classroom as well. I teach in higher ed and I read my evaluations that feedback every semester and implement what the students say. I actually give a mid-semester feedback survey. I'm not just waiting till the end of things. So I'm one of those people who does it in the middle so I can kind of change things up. But I love that your journey, you were going to go big space aerospace engineer, now you're giving back, you're educating others. That's wonderful.

DaNadia Johnson (03:31):

Yeah, I love that too. I love to hear stories about people who don't start in education but somehow find, I mean because education and learning is embedded in just about anything in any industry. And so I love to hear those stories that start out kind of different and then they're moving into this education space. So from engineering into education, what kind of emerging trends or innovations in ed tech do you find most exciting or promising that you've seen in your journey?

Stuart Grey (04:06):

Yeah, so a great question. So it's going to be around the machine learning AI space, but I've maybe got a slightly different view because part of my PhD a long time ago now was on machine learning. This was before it was cool before there were billions of dollars thrown at people with ideas Right before the cool kids were doing it, right? Yeah, exactly, exactly. I saw it before. It was cool. Way before it was cool. But it's basically mathematical models, stats, applying these things. So there's rigor there. It's not just some of the views of AI now are sort of people seeing generative AI and thinking it's a bit loosey goosey and could be all the place, but actually there's real rigor there. And I've seen that trend that changes those tools develop over time be applied to all sorts of different things, even within engineering.

(04:56):

Like AI isn't just affecting education and say office work, it's affecting everything. And you might think that I'd be a bit pessimistic now, having seen it all before, seen a few AI booms and busts, but I am massively optimistic about how fast and how much change there'll be to the education landscape in particular through ai. And we're just seeing really the first wave of these things, the first tentative steps, people throwing things out, seeing what sticks. Once we've got through a few generations of that and these generations months the way this is going, we can really start to think of how can this help us solve our very specific problems to help us do our jobs? And by that, and it again might sound strange, is doing the most human aspects of our jobs, even engineering is people think maths numbers, tools maybe, but it's communication, problem solving, collaboration, the human stuff is actually the fun bit. And the AI I think over the coming years, decades is going to remove so much of the TGM and low level work will be left with the purely human aspects that, so it might be quite a different education landscape. Education I think will look very differently than it does today, but I think it's very, very positive because we'll have so much more time and ability to focus on what really matters and what we can bring to it as humans.

DaNadia Johnson (06:27):

I really like that perspective. I don't think I've ever really heard anyone kind of say it from that angle. AI will give us more time, it'll give us more opportunity for human interaction and I really like that perspective of it.

Holly Owens (06:40):

Yeah, I was thinking along the same lines as you, Nadia, that people often see this, especially in the states. I don't know if you heard this Stewart, but as soon as chat GPT came out, New York City Public Schools banned it from their school system without even. Yep. It happened within a week or two. It wasn't that long. It didn't take them long to just ban the website and block the website from the system. So that brings me to a question for you, Stuart, since you were doing it before all the cool kids were doing machine learning stuff. Do you think that education, and we can definitely, let's look at this from a global perspective, is ready for what AI has and is about to put on the system or help with the system? I

Stuart Grey (07:30):

Would say no because for all that positive aspect, I was talking before the possibilities there.

Holly Owens (07:36):

Educational institutions are by their nature big and unwieldy and take some

Stuart Grey (07:42):

Moving. So I still work in academia as well. I still teach and we're still fighting battles from 20 years ago around. I Feel like it's the same everywhere. So it's slow progress. And that's where something like this where chat GBT comes in and people do have those initial reactions and you can empathize with Ban It, ban it. Now history tells us that's not going to work, but you can empathize with the position because there's massive change coming and the system's not ready to deal with it. And we've got to think about how we can integrate it in a sort of very deliberate way to try and support what we're doing. We're not, some people will come on and say they can replace institutions entirely, just sit at home and AI will teach you everything like that. I'm not a believer in that in particular. I think it'll just help institutions do what they're actually meant to do.

(08:38):

Some institutions have forgotten maybe exactly what they're trying to do, but actually the focus on teaching people, educating people, bringing people together to solve world's, real problems, all those sort of things this AI can really help with. But it will take time and it needs some buy-in from the few different stakeholders, institutions themselves, but also the teaching staff. There'll be some teaching staff who dive on it, jump on it and absolutely love it. But there's some who are very want to teach their way and we've got to bring everyone along. So we've got to think about how we deploy these tools in a sort of empathetic way.

Holly Owens (09:14):

Yeah, I love that perspective as well. Empathy, lead with empathy when it comes to the tools, same situation. It just seems to take forever to get through these different channels and jump through hoops and things like that. So you have a great perspective and historical aspect of machine learning and everything, but we could definitely continue down the AI road, but we want to know more about Student voice. So why don't you give us an overview, tell the audience who's never heard of it, what Student Voice is, why you founded it, give us all the information about it. We want to know more. Oh

Stuart Grey (09:51):

No, I'd love talk to, I could be here for hours. So

Holly Owens (09:54):

Student Voice is a university spin out. So it started in university. It's very important to me that it's by academia. For

Stuart Grey (10:01):

Academia, it's focused on actually solving the real problems rather than coming from some unknown vendor trying to sell you something. But we take free text comments from students and analyze 'em. So at the scale of universities, there's tens of thousands, hundreds of thousands of free text student comments come in and we use machine learning tools to categorize them, analyze 'em, group them with the aim of supporting staff to do what they want to do. So we talked about module evaluations. Before staff want to make change, they want to improve their courses. So when I was teaching those large classes over 300, get a good response rate on your survey doing everything, and then you get 200 free text comments and you just looking at this page and you start passing it. And human psychology being what it is, you look for the really bad ones. You look for the really good ones to try and make yourself feel better. There's no rigor there. It can't be because we're animals at the end of the day who are looking

Holly Owens (10:57):

To

Stuart Grey (10:58):

Try and make sense of this data. That's where it's an ideal place for something with more consistency and that can be automated, but consistency is really important. They can come in and say, look, these comments are about the lecture theater itself, these comments about the tutorials, these comments are about the assessment and show those to the staff members. And then they can go away and again, do the human thing, make the changes. They decide what's going on. Student voice isn't telling staff what to do. It's just radically shortening the time it takes to actually analyze these comments if they're being analyzed at all. Because in a lot of institutions, a lot of free text comments get given by students. So valuable time and effort by students to tell institutions what's happening. And they're frankly ignored again, through no malice, just because there's tens of thousands of them what we're going to do and they keep on coming. So we let institutions get a grasp of this, really try and tackle it in a logical way to support them to improve their teaching and learning. We have a few different aspects of what we do in particular. So we categorize things under those groupings. Like I said, we do sentiment analysis, which is really interesting of how positive or negative certain comments are

(12:11):

Within those boundaries. And you can dig into these data. Could we work at institutional levels? So we work with entire institutions generally. So we're looking at every survey run by a university across, it's everywhere, all students, all historical data. So we can get a big data set and you can start seeing these patterns of say how students are talking about assessment in more positive or negative ways. And then we can start to break that down. We can get demographic data, metadata basically with each of these comments and say, okay, this group of students has this experience, has this sentiment around this aspect of their learning journey. How can we fix that? And we can dive into that level and on an institutional level, compare those subgroupings, okay, what do the engineers think about their exams versus what do the historians think about their exams? They're very different courses.

(13:03):

You don't want to group those together. They've got very different experiences, but we can start to break that apart and that's on the institutional level. But we also very deliberately from the very start with all the institutions we work with, we share the data anonymize, but between institutions. So institutions can say, look, we're doing our assessment, the sentiment on our assessment's really high. Our students love our exams. And you can compare against other institutions say, look, yeah, the average across everyone else is X and we're x plus 5%, so we're happy. And you can see where you're below that sort of comparator. And that really helps obviously for areas of improvement. But it also helps to remove some excuses because some people will say, well, no, that's bad everywhere. Feedback on assessment, no one does that, right? But you can say, well actually you do it quite well or maybe you don't do it as well as everyone else. It's short circuit, some of those roadblocks and really allows you to make the change you want to make internally. A lot of the outputs we produce end up being business cases or cases made up and down institutional org chart to try and make change. We've got some numbers behind it, up, fire it across, fire it down to try and make things a bit different. So that's really what Student Voice does and we're growing very fast and it's a lot of fun doing this thing.

DaNadia Johnson (14:26):

Cool. Yeah,

Holly Owens (14:28):

I mean just aggregating from a holistic perspective, you're aggregating the data. There's hard data, obviously you can do evaluations with a one to five or a thumbs up, whatever. But the data, that's information that's coming in about, like you said, the sentiment, how they feel, the feedback on the longer the comments I've never thought about and putting that into categories, you just go through and you read the comments and I always try to find the good first, but that's it. And then taking that up to yourself and then taking that up another level, assuming you could probably do it by semester for yourself and then by your department or however the hierarchy is set up. So that's really cool and innovative

DaNadia Johnson (15:18):

To institution thing.

Stuart Grey (15:21):

Yeah, so the comparison to institutions is really, really fun because a lot of institutions are very big competitive, but also don't want to let anything go. So we put a lot of effort into trying to get that happening, but we could show the benefit and they want to be compared. And we are very definite, as you can imagine, there's a lot of legal paperwork goes into it's,

Holly Owens (15:44):

Oh, I cannot even imagine Comments. To get them

Stuart Grey (15:47):

Out of an institution takes a lot of work. But once you've done it a few times, you can sort of build some reputation there and trust really because incredibly valuable data to the institution. But Kant is doing nothing. So they can see that, okay, we can get the value from it if we actually share it. And then what that lets us do is once we get these sort of more and more data, and as I said, we do full historical analysis, and this isn't just modular evaluation or class surveys, this is program surveys and national surveys. We do run by governments. So we bring all that together and if we've got these institutions all coming in, we've got our uniquely large dataset. And in these machine learning classification type models, it's the amount of data you have to go into it that dictates the quality and how good your model will be.

(16:37):

So it's an institution, very smart people, institutions, they'll have machine learning experts on staff, but they're limited by, they get to see their slice of the pie and we get to see everyone slice, build these models that everyone can use. And I really do honestly see it as a win-win. So it's a really nice sort of process and it builds, and from some institutions, large institutions, a single survey can be a hundred thousand comments. And then we've got these multiple institutions, multiple surveys get lots and lots of comments come in and it just builds over time and gets better every time. It's always getting better and we train it repeatedly. We're constantly improving these things. It's just a nice ride to be on. Yeah,

Holly Owens (17:24):

I would imagine too that, what was I going to say about the institutions that don't have enough data or they have too many surveys, they reevaluate how they're sending that stuff out. Yeah,

Stuart Grey (17:36):

That's been a real aim of mine. It might seem strange, but work with institutions, work very closely with them. And some of them come to us and say, look, we've got these surveys, but they're not working their sponsor rates, single digit percentage, they're not answering what can we do? And we just work on, okay, let's remove most of those surveys. And the key thing is actually it's nothing we do, it's what we enable in that if you have a small number of surveys that people actually respond to, then response rates go and we help the institution respond to these things. We don't respond to anything ourselves, but we can support an institution doing the right thing. You talked Holly about doing mid-semester surveys. See that best practices that people know, okay, I should be doing this, I know I should be responding, but it's hard, but we can hopefully make that easier At institutional level. We do everything at institution, so it's up to institution to decide what to do, but helps them really push forward and do what they want to be doing anyway. And that's always a nice, so you're not trying to persuade someone to do something, they want to do it and they want to do it for years, and you're finally enabling them to do it. And that's a nice side to be on. I think That's

DaNadia Johnson (18:45):

Awesome. And I'm interested to know, because I know that many educators, institutions, just like with the recent global events, things that have happened in the past few years, it's been somewhat of a transition to get institutions or educators to use digital learning, digital tools to enhance their teaching and learning. So my question is, what advice or strategies would you offer educators, institutions looking to effectively integrate these kind of ed tech tools or technology to boost their learning? Yeah, that's a fantastic question. And my answer is going to be a bit strange maybe, but it's to be really boring

Stuart Grey (19:35):

And it's to

(19:36):

Try and disappear as much as possible to make an effective change. You don't want to get in anyone's way. You don't want to be some new things someone's got to learn. If you can go in there and just improve things that are broken or not happening as well as they should be so far sort of silently with no big bells and whistles, no great parades or anything like that, just quietly get it done, then there's no need for the staff to know what's going on. Even all they know is that for their class or their program or their department, they get some labeled text. They don't care the backstory, they just want to get stuck in there. It is just silently get their aim. And maybe this is just how we approach it, but to be sort of a boring foundation. We plumbing, getting things working.

(20:22):

And that could be using cutting edge technology, doing very boring things. But I think that's a good niche to be in because you can see already some ed tech solutions coming with a big bang and oh, if everyone moves to this, it'll be amazing. And it would be. But as we've talked about institutions, getting everyone to move to something as an active thing is very, very difficult to do. But if you can go and a busted process, it's a really easy win. And that'd be my advice of how to approach these things more generally. Think of what is currently really difficult or impossible and see if the tools can help do that rather than replacing what already works in some sense.

DaNadia Johnson (21:03):

I like that perspective, kind of easy implementation. Nobody wants anything that's going to be too much to learn, too much to implement, too much to do. I like that. Or They have to go sit through a required training. People hate that. The silent tool that's going to just enhance without the educator having to overwhelmed, be overwhelmed or think too deeply about it. I love that.

Holly Owens (21:26):

Yeah. I don't know about you Nadia, but some of the best ed techs I've used over the years just kind of fade into the background of the classroom

DaNadia Johnson (21:32):

And

Holly Owens (21:32):

They just kind of become a part of, like you're saying, Stuart, like a part of the routine and they're just there and they're fixing something. So it's really, that was a really insightful perspective.

DaNadia Johnson (21:42):

It was. Yeah.

Holly Owens (21:44):

I like

DaNadia Johnson (21:45):

That

Holly Owens (21:46):

A lot. Well, we have talked about a lot on this episode and we're wrapping up here and we'd love to have you back to give us an idea of what's going on at Student Voice, but we have two final questions we'd like to ask before we let you go. So the first one is, is there anything that we missed, anything else you want to tell us about Student voice, get it out there. And then we also want to know from you, what is the future of ed tech or education? What is this all looking? So we talked about some trends, but we want to look more into the future of what that's going to be. So anything we missed and then tell us about the future.

Stuart Grey (22:28):

Nothing, but that's a great chat, but again, we could probably chat for hours now. I'll have to jump Back. Yeah, I have so many more questions.

Holly Owens (22:35):

If you'll get an automated thing from our scheduling, there's six months from now, so please schedule a Follow-up.

Stuart Grey (22:41):

I'd love to. Love to, love to. Yeah. So it's really, and I could probably answer both questions at once. What I'd like to reemphasize, not anything else I'd like to share apart from if institutions are interested in maybe improving their systems around this, obviously drop the line, we'll have a chat, but

Holly Owens (23:00):

Oh, this all going to be in the show notes, so they're going to know where to find you, Stuart.

Stuart Grey (23:05):

What really gets me going is that human aspect, and that's what I like to reemphasize is that sometimes people can look at things that automated process as removing people, but I talk to people whose previous part of their job is manually labeling comments at universities and they can't wait to get rid of that because the other part of their job is making the change, actually improving things which they love. So we are freeing people up. So it's that really that freeing people up to do the very human thing, the teaching side, the communication side, trying to improve things collectively. And that's really what I had like to, and in terms of what the future looks like, the technology is coming, it's a tidal wave, let's say once it washes up and washes back and we want our time to sort of catch our breath a bit in a few years time.

(23:56):

I think we do have to think about what is important in our education. And as an engineer, it is really those human aspects and one of the most technical subjects there are, but still the human stuff that interests me. And we've got to start thinking now about what does a technical subject look like when you could have assistance to basically take care of all the technical stuff, the historical, the math stuff, which as engineers have to know, but then what does engineer have to know? Then they have to know maths but not memorizing formulas I used to have to do in school instead. It's like mathematical thinking, logical problem solving, abstraction synthesis, taking things from one place to another. All these really fun skills, which are mathematical skills but not maths in terms of what we use. So we really have to start that change now. As I said before, education is very slow moving, but we want it point it in the right direction of helping the most people and these tools will help us that, but we've got to work together to guide it. Really.

Holly Owens (24:53):

Oh my gosh, 100%. Totally agree with that. And I love that there's classes in some of our institutions called human computing interaction, so obviously people have already been thinking about the human aspect of how things are going to interact. And what you said before, just touch on this a little bit, is that it's not replacing the human it's coming in and helping us get more time back. I feel like what you're saying is there's so much manual labor, especially in academia that goes into certain things and we need that time back.

DaNadia Johnson (25:30):

And I love that perspective because I just feel like I haven't heard that perspective when it comes to AI yet. Yep. Well,

Holly Owens (25:39):

Stuart, we can't thank you enough for coming on at Ed Tech and chatting with us about student voice and sharing all your AI knowledge and just everything that you've been doing and we really appreciate what you're doing for the community and we had a great time.

Stuart Grey (25:53):

Oh no, thank you very much. It's been an absolute pleasure to chat with you guys and hopefully come back on and see how it's all going, where my predictions are right and where they're very wrong. It's

Holly Owens (26:02):

Always

DaNadia Johnson (26:02):

Fun to

Stuart Grey (26:02):

Chat. Yeah,

Holly Owens (26:04):

Absolutely.