How big will the AI in Education revolution get?

00:01.82

Rishi Raj Gera

Hello everyone, this is Tech in EdTech. In this podcast, we discuss technology that powers education and improves learning for all. Welcome to today's episode. I'm your host Rishi, from Magic EdTech, and our guest for today's podcast is Josh Jarrett, SVP of Strategy at Wiley. Josh is an experienced innovator and executor at the intersection of higher education and employment. He has built products, programs, and relationships that have helped hundreds and thousands of people successfully earn degrees and secure meaningful work. Josh, thanks for joining me, and welcome to today's show. 

00:39.36

Josh Jarrett

Thanks for having me Rishi. I'm really glad to be here.

00:44.78

Rishi Raj Gera

Josh, why don't we start with having some background about how you got into education and your journey since then?

00:50.29

Josh Jarrett

Sure, I actually started my career in management consulting and product management during the original dot com bubble, back in the late 90s, early 2000s, and really switched my career to non-profit consulting and used that to talk my way into the Bill & Melinda Gates Foundation in 2006 and the charge there was how do we reduce inequity in the US in the most cost-effective way possible and so as part of the team that launched our postsecondary success strategy focusing on helping more low-income students, students of color get access to and succeed in and earn a credential with labor market value beyond high school. And so I led the innovation and technology portfolio there for 7 years and then subsequently I have worked in a number of organizations at that intersection between education-employment. I co-founded a venture back startup, I spun an education benefits company out of Arizona State University. I worked at a university in Australia on their short course programs online programs, I created a nonprofit during covid, top covid impact to workers get short-term rescaling and in-demand careers and then two years ago I joined Wiley to lead their ah lead our strategy which means looking at the future strategy of the organization, threats, and opportunities. Our M&A strategy and I've built an internal innovation team as well during that time.

02:15.29

Rishi Raj Gera

Wow, That's, that's truly interesting. You know I see your entire remit has been higher education and you have been very instrumental and pivotal toward you know this domain.

02:25.84

Josh Jarrett

Yeah, most of the work has been in higher education and really the intersection between higher education and the workforce. So. It's really about how we help people fulfill their own individual potential and how we help them launch successful careers. And how society gets the most out of all of the potential that we have distributed across our population.

02:49.30

Rishi Raj Gera

Interesting. So, Josh, as you know that artificial intelligence has been pivotal and its promises are big with respect to you know the higher education, including personalized learning, improving student support, and more efficient grading and data analysis. So, however, like any technology, AI is not a panacea and right now there's a lot of uncertainty surrounding its use in education. Can you give us your perspective on this?

03:15.38

Josh Jarrett

Sure. So both personally, and at Wiley, we are absolute supporters of the responsible use of AI in education. We think it's going to unlock some real benefits to learners and to organizations. Ah, but we also recognize that there's some serious threats that that brings so we've got to be thoughtful. We've got to be smart about how we use humans to set guardrails and to do quality control. But I think we've got some real analogies to build off of. If you think of the calculator, if you think of the internet, if you think of Wikipedia um, each of those has been a new shock to the system, a new technology. We got very initially alarmed, what is this going to mean? And we worked our way through it and we got to an equilibrium point and I think the same thing is going to be true here. There's so much buzz about ChatGPT and Generative AI but one way to think of that is a really, really, really good Wikipedia um, and so I think if you think of some of the history that we have, we've got tools that we can use to manage through this.

04:22.88

Rishi

Does that mean that you know AI potentially could help our education systems which are inching toward that holy grail of building a student base, particularly in the workforce readiness market?

04:34.26

Josh Jarrett

So I do think there's a lot of opportunity and to be clear, I think there's two big opportunities that AI can help us in that holy grail of getting students' workforce ready to succeed. There's a big piece around learning and there's a big piece around talent discovery. So, on learning Benjamin Bloom of Bloom's taxonomy did research in the 80s called The 2 Sigma Problem. We have known since the 80s, how to improve student learning by 2 sigma or 2 standard deviations. The answer is an individual tutor and mastery-based learning. The problem is we've never been able to afford giving every learner a personal tutor. So. It's not actually a limit of our knowledge of what works. It's our inability to afford it. And so the potential of AI to create truly personalized learning experiences where you can ask a Generative AI engine say, explain this complex chemistry concept to me like you're my older cousin. Um, you know who you know, who has this background or that background, right? And you can get a structured tutored experience back around. Well here's the concept, here's why I chose that ask me a question, I'll explain it, and so I think we have to harness that but there's a real opportunity to create much more personalized learning. And then on the set on the other side is of Talent Discovery. Yeah, these big patterns allow us to connect potential with opportunity. So one of the startups that I helped co-found that I mentioned was called Koru. We built predictive hiring engines for companies to do early career talent hiring. So we would look at soft skills assessment plus signals in someone's background and could predict with 20 to 50 percent more accuracy if someone was going to be a successful hire than the interviewers themselves and so I think that we're going to be able to connect more people to better opportunities through some of these AI engines as well.

06:37.97

Rishi

I totally agree. You know into this whole lifelong learning practically you know AI is definitely instrumental and in fact, to the audience today who've been listening to this podcast. I think more are aspirational workforce, who are definitely seeing it both as a threat and an opportunity and you know all your insights are definitely going to be meaningful for them. So, moving on to my next question, when Chat GPT was released, there was a lot of uproar from every sector, especially whether it's banned due to plagiarism but you can't stop advancement in the technology, just find ways to harness them. In what ways do you see education doing this? What's getting easier?

07:21.50

Josh Jarrett

Yeah, I think that um I think that we do have to embrace the technology with smart guardrails and I think those guardrails are really important and those guardrails are going to vary depending on the stakes of the moment and you know in this specific context. But generally speaking, technology that makes the human condition easier will almost always win. If this makes someone's life easier, they can get something done faster with less effort. You know water runs downhill don't bet against water running downhill and I think this is one of those moments where um, where the, you know the water, the dam is broke the water's running downhill. We've got it. We can redirect it, we can steer it. Um, but it's going to keep keep keep running downhill. So I think there's some real specific opportunities with AI and I'm particularly, I'm thinking about generative AI right now. I think there's three, one the creation of educational content gets much easier, faster, and cheaper. We're doing experiments at Wiley right now in creating assessment questions, at creating supporting videos very quickly to support text and other materials. So I think you can create a wider range of content, so it reaches a wider range of audiences and at a lower cost. So content creation is certainly one of those. The second is creating more personalized learning experiences like I talked about with these kind of intelligent agents or personalized tutors that can really help people solve that 2 Sigma Problem I mentioned earlier and then the third is a really big one which is much of education is currently and even more so in the future going to be about motivation, the motivation to learn. So how can we use some of these tools to do that? Think of this way, I'm sort of struggling, I'm burned out, I'm tired, It's late at night I'm studying, and you know here's a button that says create a picture of me in a job that uses this content this thing I'm studying right now and then suddenly there's an image. Literally, It's me doing a job in the future that I'm happy and I'm liking and I'm using my engineer um, geometry in an engineering application building a building ah a bridge and that's interesting and that motivates me and it puts me back to work. So I think that's an example outside of just the learning moment where there's ways that we can help students connect opportunities that will excite them.

10:56.53

Rishi

Yes, so if you can also talk about certain risks as ah, you might have a better view from a content provider perspective. Where do you see you know the industry going in and taking those perspectives?

11:11.20

Josh Jarrett

Absolutely! So there are so ah, there are 3 risks that I would call out, one, shortcuts in the learning process. You know it's just that we as humans look for ways to get the amount of work done with you know the shortest time possible. But there's also saying that it's the work that does the learning and so if I'm looking for ah you know a shortcut? Um, am I if I ah you know, actually shortcutting the learning? But for example, if I'm about to do ah I've had to a long paper on The Weimar Republic and how that ah you know led to conflict in Europe. I might go to Chat GPT and say, can you generate an outline for me on this topic and that might help me create an outline, and that might allow me to get started working on that paper. Is that cheating or is that just me kind of you know helping make my life a little bit easier? Getting started? Um, we have to figure out where the rule the line is there but certainly, those shortcuts might shortcut my learning. Then there is outright cheating and plagiarism. It's going to be that much easier to say no go ahead and write me that whole paper on the Weimar Republic. Um, and so we're going to have to be really intentional and thoughtful about that. Um, ah, the third one I might mention is that we risk confusing learners. This technology and Generative AI, in particular, is a people pleaser. It will do whatever it takes to prove whatever point you ask it to prove even if it's not true, these are called hallucinations. It'll say here's, here's a justification for the point you want to make but it may not actually be based on fact or citations or anything like that and it also carries the biases that and that that were built in you know from that original data. Um, and then I also the last thing I would just mention is educational inequality. To an extent, these tools aren't free. Um, and certain people get access to them first or people are able to learn the prompt engineering right? The way to manipulate these tools to help them that I think we can also exacerbate in educational inequality.

13:17.47

Rishi

That's so awesome of a view. How do you think ah you know educators will see this innovation? Are you thinking that they will be accepting this out or do you foresee that this technology will be seen as user-friendly or more complex than its predecessors?

13:36.25

Josh Jarrett

So I think HigherEd and Education educators, in general, have been relatively slow to adopt new technology. We know there's still some VCRs in some classrooms. Um, but I think that we're gonna have to embrace it and face it pretty quickly and we're gonna have to set some guardrails that define here's where you should be using these tools and here's where you can't um and I think it's going to change the way that education happens. In that, it may be the end of graded homework. Why have homework when I send you at home and I can't tell what you've done? What the machine has done? So I think we might have less graded homework and more and graded formative assessments and it's going to actually raise the stakes on fewer high-stakes assessments. Remember things when we were growing up like closed book test, pop quiz, oral reports, oral presentations? Those are all tools that isolate the learner's brain from the machine's brain.

14:41.65

Rishi

Yup.

14:48.45

Josh Jarrett

And I think we're going to, you're going to see things going back to creating those types of experiences that help isolate what that learner has actually learned but then in the formative moments when you are doing your homework I think it creates a real opportunity to help create better formative learning experiences. So better form of learning experiences with more controlled higher stakes summit of assessments.

15:16.62

Rishi

That's awesome and do you think these thoughts could also portray on bridging in the gaps in equity learning?

15:24.78

Josh Jarrett

I think it's going to It's as many I think it will have both a positive and a negative effect on equity and learning. As many of these technologies do, they create opportunity around access and around personalization, and around making up for ah, resource limitations and they create the opportunity for those with resources and knowledge to run further faster. Um, so for instance right, think about test prep courses, right? SAT prep courses, Princeton Review and Kaplan, and all those ah right that those have allowed certain people with access to take 2 or 3 of those classes before they take the SAT and perform better. But now we have digital you know we can. You can take those classes for free or near free online now and so more people are able to to access those. Um, so the idea that we could give every learner around the world a private personalized tutor is an amazing aspirational possibility but the reality is we will get to the solutions that accelerate the haves first before we complete the promise of helping get this technology ah into the hands of everybody else to close the gaps.

16:47.63

Rishi

Interesting and within your own role and your and your enterprise are there any upcoming signs of progress in this field? Be it like you know some platform providers, some tool providers that you find exciting or those are catching your eye at the moment.

17:04.40

Josh Jarrett

Yeah, I, you know, look I think the first application of these is actually going to be in content creation for, so why he's a publisher right? We are a trusted curator of knowledge and that comes with a very high bar in terms of making sure that people can trust the quality of the content um and 60% of Chat GPT comes from the open internet the content from that. So It's pulling in a lot of things that aren't validated, aren't curated, and aren't authenticated. So I think there's a real big opportunity to save time and money in creating high-quality curated content. But it's still mediated by humans and so we're doing a lot at Wiley right now to increase the speed agility and cost efficiency of content creation that still has the quality bar that you can trust. So that's the first easy application for us. I think the second one is then more around these personalized learning experiences so that we can. We can start to put these tools in the hands of learners and teachers so that they can interact better. And then the third is this longer-term question of creating new ways to access information and content and connect to talent that that that builds off of these models for instance on our on the research side. We do a lot of work at Wiley on research publish academic research publishing. We're using AI models to find peer reviewers, to find co-authors, to find experts around the world in your particular area that you might want to write a collaborating write a paper on or who ought to review your journals which is creating a much more diverse set of researchers working together and so these new networks that get formed are really exciting to us.

19:03.84

Rishi

Interesting and in what ways should businesses and institutions in education prepare so that they are prepared for the adoption and usage of AI? Are you seeing ah your your data policies or data governance frameworks being changing or what kind of guidance do you have for businesses here?

19:22.35

Josh Jarrett

So, the worst thing to do I think is to do nothing because people are going to start experimenting and ah you know going to be confused or they're not going to want to do anything because they're not sure so I think be very clear about approved use. Here's what we encourage you to do, don't ban it outright but define the approved use and give training on what are the risks. Here's what you need to know, this content that comes back it has hallucinations in it is not necessarily fact-checked and validated. It'll tell you it's a people pleaser. It'll tell you what you want to hear ah, think here's what to do about ethical or moral considerations in what you return here. Here's how this data that you put into these models is going to be used. Um, so don't put our proprietary content into a request on this particular. Ah ah, large language model prompt engine because they're actually taking that content and building it back into their models. Different providers don't do that so setting up the clear guardrails and um, use cases and protections is really Important. I think the second thing is thinking about where are the stakes the highest right? Where the quality of information must be accurate, where the quality of the assessment must be accurate, and try to make sure that you can create um the type of of learning moments or assessment moments or what have you that are protected from ah from some of these risks. And then the last thing is to think about hey use 10 or 20 percent of your time to look for the disruptive business models. Ah, there could be some real changes to how your customers are going to access information or going to find networks or going to generate content on their own and that's going to create potential for new opportunities. But it could also radically disrupt some business models that are out here now. So it's the thing that you know comes at 4 AM in the morning when you're sleeping that disrupts your whole business model. You got to figure out a way to spend enough time to be looking for those.

21:30.80

Rishi

Great, great, and in our show today you know we definitely have a lot many edtech leaders and educators listening. What is their responsibility? What do you think they owe to the society today?

21:44.89

Josh Jarrett

Yeah, look I think, in the history of human knowledge acquisition and learning up to this point in time we've been standing on the shoulders of giants who are human. But we're at a seminal moment where the next giant on whose shoulders we're going to stand will probably be a machine. For the first time ever, we're building our knowledge and our but on top of a machine's input and knowledge that's exciting and that's scary at the same time. It's a milestone in the history of humanity and so we need to really thoroughly explore the pluses and minuses of this new development because it's not going to go away. We can't put the genie back in the bottle. Um, but we got to figure out how to do this not just in like what's the quickest way I can make my edtech company figure out how to make a buck by saying I've got a new you know, Generative AI application but we've got to figure how to use it responsibly morally and ethically and that's going to be something that each individual, innovator, entrepreneur, and edtech companies got to think about, but it's something we need to think about as a community and as as a group of edtech innovators because it's a lot and it's moving really quickly and we got to put our best brains on it.

23:03.75

Rishi

Josh, I would like to thank you for joining me today for the latest Tech In EdTech podcast. We really appreciate your insight and look forward to you and our audiences joining us in future podcasts. Have a great day ahead. I believe this was a meaningful session and I'm sure you know all of our audience will definitely value from it.

23:21.69

Josh Jarrett

Rishi, Thanks for having me. I really appreciate it. It's exciting, exciting materials and I look forward to listening to more of your podcast coming up.

23:27.29

Rishi

Thank you, Thank you.