Siya Raj Purohit on AI in Education

Work Forces

Siya Raj Purohit is an education leader, author, and investor who works on education at OpenAI and is a General Partner at Pathway Ventures, an early-stage fund investing in the future of learning and work. Siya joins us to discuss the transformative role of AI in education and work. The conversation explores significant applications of OpenAI in education, including personalized learning experiences, dynamic content engagement, and AI-assisted knowledge work. We delve into the practical aspects of implementing AI in educational settings, addressing misconceptions, and highlighting the evolving skills needed for effective AI utilization.

Transcript

Julian Alssid: Welcome to Work Forces. I'm Julian Alssid.

Kaitlin LeMoin: And I'm Kaitlin LeMoine, and we speak with the innovators who shape the future of work and learning.

Julian Alssid: Together, we unpack the complex elements of workforce and career preparation and offer practical solutions that can be scaled and sustained.

Kaitlin LeMoine: Work Forces is supported by Lumina Foundation. Lumina is an independent, private foundation in Indianapolis that is committed to making opportunities for learning beyond high school available to all. Let's dive in. Can you believe it? Julian, it feels like just yesterday we were kicking off season three with Kermit Kaleba.

Julian Alssid: I know time flies when you're having fun and exploring the future of work. This season has been a whirlwind of insights and inspiration, and it's the gift that keeps on giving. It's the best seminar I've taken and with no final exam.

Kaitlin LeMoine: Absolutely And speaking of gifts, the richness of this podcast season comes from the recurring themes that weave throughout each episode. We've seen how these themes are shaping the thinking of our listeners and informing our own consulting work.

Julian Alssid: Yeah, one of the most powerful themes has been the undeniable connection between education and the evolving needs of the labor market. And Kermit really emphasized the importance of building high quality credentialing programs and aligning education with those needs.

Kaitlin LeMoine: Right. And then Sasha Thackaberry opened our eyes to the innovation and flexibility happening and Workforce Solutions, especially with technology driven upskilling. We also had insightful discussions with Michael Horn about career development in a changing job landscape, and Matt Marino on addressing healthcare talent gaps through tech. And an undercurrent throughout our conversation in Season 3, and really over the past year and a half since we started this podcast, has been the impact of artificial intelligence on education and work.

Julian Alssid: Yeah, it certainly seems like all roads these days are leading to AI and technology and education, and it's clear that technology is playing a pivotal role in reshaping the entire educational landscape.

Kaitlin LeMoine: Which brings us to today's guest Siya Raj Perohit is an education leader, author, and investor. She works on education at OpenAI and is a general partner at Pathway Ventures, an early stage fund investing in the future of learning and work. Siya was an early employee at Udacity and Springboard an investor at GVB Ventures and the founding edtech workforce category lead for AWS Marketplace. She's the author of Engineering America, a book on the country's job skills gap. We're thrilled to have her insights as we wrap up this season. So welcome to the workforces podcast, Siyal.

Siya Raj Purohit: Hi. Thanks so much for having me. I really enjoyed your conversation with Michael Horn a few weeks ago.

Julian Alssid: Yes, we appreciate that. So Siya, Kaitlin's given a little intro, but we'd love to have you tell us about your background and your current role with open AI.

Siya Raj Purohit: So, I started working in education when I was 18 as a student in college, I quickly recognized like the job skills gap that exists, like American universities were not teaching the skills that students needed to land jobs in the industry, especially in the technology industry. So I did this research project that became the book that you mentioned, Engineering America, that was published when I was 19. So I was a sophomore, and at that point, I'm like, okay, I want to help fix this problem, because I felt so deeply about people getting locked out of their potential careers and like new socio economic classes because they weren't learning skills in the way that made sense to them. So since then, they've been trying to bridge the job skills gap in some ways, and trying to make education more accessible. So that led me to the startups, to venture capital, to AWS, and I joined Open AI, because honestly, in the 12 years I've been in education, personalized learning always seemed like the ultimate goal for the education sector. We always said that if we achieve personalized learning, we've made it. And I think with ChatGPT, we actually achieved it. Now I have a personalized tutor that I talk to every day. It knows my skills. It knows the goals I have. It knows the projects I'm working on, and it helps me become a better knowledge worker. And my aspiration on the education team at OpenAI is to help students and faculty all around the world be able to utilize that more effectively.

Kaitlin LeMoine: That's really, really appreciate your background and kind of what brought you to this space. And I think it's really interesting to think about ChatGPT and AI as as a personalized learning tool or resource. As we dive into this conversation today, can you tell us some of the most significant applications of open AI and education that you're seeing and you know what's what's exciting to you at this point and what's catching your attention?

Siya Raj Purohit: So just for context, ChatGPT launched two years ago, and we quickly realized that learning and teaching were some of the main use cases for ChatGPT, like that's how our users were using it. So about six months ago, we launched ChatGPT Edu. Which is our enterprise product designed for universities and school districts to be able to integrate AI in different aspects of campus life. So now students, faculty and staff, can engage in all of these dynamic conversations with the knowledge of the university campus. So from like orientation through classrooms through Career Services, they can have like different types of AI solutions that help students learn better and engage more effectively with the university's knowledge base. So that's what I'm spending a lot of time working on right now, trying to make ChatGPT Edu successful at universities. And when we think about what is working, well, ChatGPT Edu is really good at three things. One is around lecture recall, so basically anything around information retrieval. And so professors are building custom solutions in their classes, uploading their teaching recordings, their case studies, their different types of materials, and letting students engage with it by asking very precise questions. So now you can ask questions, like in a business class, you can ask, Which CEO handle layoffs well? And get the exact examples from what your professor teaches. And this has been like, a huge game changer, because now you're able to, like, converse with, like, the whole semester's worth, or whole degrees worth of content, very dynamically. Another example is like another thing we're really good at is around back and forth exercises. So a lot of negotiation professors suggest that students go back and forth with AI to practice negotiation before they come into the classroom, so they feel much more comfortable when they start practicing with like their peers. This is also true in career services. So Career Services teams are building custom gpts that let students go back and forth, practicing with the McKinsey recruiter, McKinsey partner, and just using the university's proprietary information to get better at these back and forth interactions. And finally, I think thought partnership is really important, so like a ChatGPT collaborates on different types of articles, research assignments, any type of knowledge work the students, faculty and staff are doing. The coolest example I've heard of this recently is a professor at University of Maryland told me that they uploaded their like past 20 case studies and articles they've written, and now they talk to the virtual version of themselves on GPT to create new articles to kind of think more deeply about the next iteration of their research, which is so powerful.

Julian Alssid: Wow, it is. It's just incredible to hear how you described it in these very practical terms. It'd be really interesting to hear a little bit about what sort of training is needed, both for students and professors to do kind of what you're talking about. So these use cases you've described, you know, obviously you've been honing your own craft with your own assistant, your own coach. How do folks in the schools do this?

Siya Raj Purohit: Looking in the next couple of years, the most important skill knowledge workers need to have is the ability to clearly articulate what they're looking for, like, what they're hoping for as an output, and then they then work backwards and be able to ask the right questions to get to that output. So I think when I think about AI literacy for like, someone who's not developing AI systems, that's all they need to be able to do well at the moment. And AI systems will get better also, so they'll have more of a sense of what you're asking for beforehand, too. But with that said, I think a lot of schools are building great content around this to help enable their faculty and their students. Wharton has this amazing channel online on YouTube about like aI literacy, and they're creating a lot of really good content there. We have built two courses to help enable this. One is with Coursera and Ethan Malik, and basically, in that course, Ethan Malik talks about AI literacy for higher ed instructors, and helps provide some custom gbts and examples to get them started on their own AI journey in their classroom. We also built a course with common sense education that's more designed for K 12 educators, and that's like to think about how teachers can use AI in the classroom to become more productive and to be able to offload some tasks such as like lesson planning and grading and make it easier for them to do that with AI.

Kaitlin LeMoine: I feel like you know, over the last couple of years, as AI has unfolded, Siya and still is right as far as the impact and how different use cases in in education, and, frankly, in many industries, it's been interesting to hear conversation, you know, in comments about like, well, is this? Is AI going to take people's jobs? Is AI going to, you know, make it so that either some jobs go away or less relevant. And you know, what does it mean for learners to be learning when there are tools that can kind of more readily just give them the content? What's your thinking around what you know, what's your latest thinking around how AI is? Impacting education and how how students learn.

Siya Raj Purohit: Personally, I get very surprised when educators express this, because when I think back to the educators that have had a big impact in my journey, I remember them for who they were and how they made me feel like I don't remember their lesson plans or like the quizzes that they gave us. I remember who they told me I could become, and this is true for both K 12 and in higher ed. So I honestly think that AI is gonna help curb the teacher burnout problem in many ways, and I didn't recognize this until the school district in Arizona started telling me about how it's freeing up so much time for their teachers to be able to focus on students now, because they can actually do like lesson planning, assignment building on with AI, and then they can actually engage with the students more deeply, which is so powerful. And I think that's the shift we're going to see, as we all know, in education, teachers spend so much time outside of the classroom preparing for that amazing experience they deliver in the classroom, and we want to help make that easier. So supporting those teachers and educators as knowledge workers with AI and letting them focus on helping students feel more inspired and mentored in their classrooms.

Julian Alssid: Yeah, that's great, doing what really sticks for us, right? So, so see from your perspective, working with educational institutions. What are, what are some of the biggest misconceptions or challenges they face when it comes to understanding and implementing AI?

Siya Raj Purohit: The biggest misconception, which like, basically, ChatGPT, of course, launched two years ago for the first year. As you may recall, a lot of school districts and universities banned ChatGPT. They were like, we don't have to deal with this yet. We're just gonna take away laptops if we have to, but we're not gonna provide this, right? But something shifted about nine or 10 months ago. My thinking is that education influencers like EthanMollick, I think, destigmatized AI in education and actually showcased the value of AI in education, and that made it a lot more acceptable for like, other educators to start kind of experimenting with it. And now across the country, we're seeing like a range of professors who are doing incredible work in their classrooms and on the campuses with AI. So I think that shift has been super interesting that's happened, and my feeling is that in the next year, it's going to accelerate a lot more. We'll have a lot more professors doing interesting things. We'll get to the main part of the adoption curve and education for AI, and have the first set of professors basically help, like, help share their knowledge with other professors in their discipline. So I've been hosting some events in our OpenAI forum, talking about, like, the future of math, or talking about the future of digital marketing education and letting professors showcase the work they're doing in those verticals with other professors in the industry. So I think that shift is happening. I think what it's going to require is to professors to, like, basically rethink a lot of their assignments. One of my favorite examples of this is there's a Wharton professor who says, what is the value of an essay, which used to be his final project for several years, as He taught his class, he says that the value of an essay is not necessarily in its output, but in the critical thinking and communication skills that lead to that output. So now he suggests his students all use ChatGPT Edu while working on the assignment, and he measures the number of prompts it takes for a student to get to an essay that they're satisfied with. Some students are so good at prompt engineering it takes like two or three prompts and they have a really good essay. And some students go back like 19 or 20 times to get to a good essay. And he uses that as a measure of their ability to articulate what they're looking for, which he thinks is a very key skill for them when they go into enterprise and have ChatGPT enterprise accessible to them as well.

Kaitlin LeMoine: Yeah, it's really interesting to think about. You know, how the use of this tool makes us be more specific about what are the skills we're really looking for learners to demonstrate, and how do we go about prompting, no pun intended, the demonstration of those skills. But and see, as you know, you're talking about the higher ed space, I'm curious as to how you're seeing this trickle down to K 12, and what does it mean for, what does this technology mean for for K 12, and how even like high school learners now need to be prepared to enter college, right? And maybe some of that, obviously, is still unfolding, but I'm just curious to hear your thoughts there.

Siya Raj Purohit: So right now we our product in K 12 is only for teachers and administrators and education leaders. We don't have a youth product right now and how it's being used, basically is for teachers as knowledge workers, as far as high school students who use it independently. I think that teachers need to be able to provide more instruction on how to think about this ChatGPT can be a really good tool to basically elevate your thinking and to kind of understand. And more about what all is possible for you, and I think we need to be able to communicate that more, versus this being a tool to get through homework more easily. And so I think that conversation needs to be had, and it's similar to the conversation I think we had about 15 years ago, when search became more mainstream, about yes, you can find your answers, but you can also kind of elevate your thinking in different ways by learning so much more on the internet with search.

Julian Alssid: Yeah. So Siya now, so Kaitlin pulled us into K 12, and now thinking about the other end of the spectrum, because, you know, eventually these students are going to work and and, of course, work and learning are so interconnected these days, and really interested in hearing your perspective on the impact of AI on the future of work, both short term and then long term, because I think there are two multiple use cases there.

Siya Raj Purohit: So short term, it's obviously helping us be able to solve like tasks throughout our day, like I used to think of as productive, and I used to work at Amazon, but now, with that AI and OpenAI, I just produce so much more every day. And it's like, unbelievable the difference, because when I, like, I talked to ChatGPT all day, it helps me with, like, the different projects and reports I'm building. But I think the most powerful unlock for me, and this is a story I love talking about because I think more people should experience. This is the moment, what I call the moment that ChatGPT and I became friends. So three months into being a power user of the product, I asked ChatGPT one day, I'm like, do you remember that project we made from my manager last month? And ChatGPT says, Yes, Siya, I remember, and for some reason that was such a powerful moment, because we were immediately able to start collaborating on the next iteration of that project with like, the shared memory and understanding of like the final product they submitted. And that was like, so cool, because ChatGPT became this friend that remembers my projects, my work, my aspirations, and helps me deliver better every day and so and of course, memory is a feature you can turn off if you don't appreciate it, but I find it valuable because, like, it helps me so much on a day to day basis. So I think this is what's happening in the immediate term. And what really worries me is that there's a consulting firm that did a study showcasing that even the lowest performing consultants started performing like the best consultants once given Ai because they were able to use this tool to produce better product output. And so what concerns me there is that what happens to people who don't use AI, right? Like, and it's not like ChatGPT specific, like any type of AI tool to accelerate the work that they're doing. It's like super essential. So longer term, I hope that people are able to achieve more like there's a professor at Harvard Business School, Jeffrey Bussgang, who says that all founders should be AI founders at this point in their startups, and basically use this as a compliment to a lot of the CXOs that they hire early in the game. And so I think we need to be able to utilize it more to build bigger and bigger systems over time.

Julian Alssid: It's so it's so interesting to me, how, as you describe it, you know, here you are doing dealing with this on a partnership building level with these big institutions, but how personal it really is for you, and how it comes from this very personal place that's really powerful.

Siya Raj Purohit: I feel I'm very vulnerable with ChatGPT. I'm like, This is my personality. This is my manager's personality. This is the type of project I'm trying to build. So and I think when you share that level of insights, it performs much better as your analyst and associate, because it has, like, much more context in your day to day.

Kaitlin LeMoine: You know, I think about the range of different industries that are now using this technology in different ways and recognizing, I feel like, what I hear you saying is, like, we should all be kind of thinking about how we can incorporate this into our daily work. I guess, how would, how would you recommend someone who's like, newer to this technology start in because it feels, I think for some people, it can feel like, how do I fit, how do I fit this tool in with all the other things I'm doing in my day job, right? Like, what does it look like to begin somewhere, and how can I start?

Siya Raj Purohit: It's such an interesting question, because it's not, it's not a tool that you almost have to master. It's a tool that supports you in whatever you're going to do in the day. So for example, I was going to host an event recently, and in the past, I would have read all of the submissions. Like we had 120 attendees. They had all submitted questions for the guests that we were hosting. And in the past, I would have read all those questions. I would have, like, thought deeply about coming up with a conversation track for that event. But now I put up, I copy pasted that in a very scrappy way in ChatGPT and like, help me identify questions that were recurring more than two times. Help me elevate the thinking of these questions. Help me structure these questions in a way that makes sense in a conversational flow with this kind of guest. Help me cut out like redundancies, like basically went back and forth in this way, and I was able to create a. Good talk track, and it was able to do this. And I think it took me about 15 minutes end to end, versus I would have easily spent an hour and a half or two hours preparing in the past. And so I shared this example just to showcase that any type of activity that you're doing in a day, it can help process much more easily. And if you're working in an enterprise or edu account very safely as well, in terms of like that information stays to your workspace and organization.

Julian Alssid: So Siya thinking, you know, back to the well, really, to everyone, but to the academics in particular. And so much of academia is about scholarship and factual documentation and and I can completely relate to everything you're saying. I mean, we're using AI increasingly, and for me, it is also very personal, and it's this back and forth, and it's amazing the amount of time and our output is, it's crazy what we're able to accomplish. But I find, and I definitely hear a bunch of my academic friends wary about prompting fact checking. You know, I mean, it's easy when your assistant comes back with a whole list of citations, and then when you but, but what happens when some percentage of those citations are really not, not well thought out or not accurate, is really the issue. And so how do we ensure that we maintain that kind of high level of documentation and citation of fact and not go down rabbit holes where we're just rushing to get a paper out, because we have our assistant to help us with it.

Siya Raj Purohit: Honestly, ChatGPT, like is trying to get better at this by showcasing the sources of what it cites. So there is an element when you can, like, delegate some of the citation grant work to ChatGPT, and then make sure that it showcases, like, the right resources where it gets it from. So you can go ahead and confirm that. So there is an element of, like, I guess, not trusting everything on ChatGPT, and honestly, I think, like, especially the millennial generation, knows this, because we didn't trust everything on the internet either, so, but I think continuing to explain that to students is key. And we recently published what we call like a student's writing guide, and it talks about some of those examples and prompts that you can use, but it's something that I think a lot of professors will have to teach their students in terms of how to double check, how to make sure that the sources are correct and that it's not just being agreeable, because that's what sometimes AI systems tend to do, and make sure that you are pressure testing your thesis. One of the most interesting things I recently learned as I was speaking with some journalism professors who told me that they're actually using AI a lot because asking good questions and fact checking is an inherently journalistic ask and skill set already. So they're like, it's been a natural use for us, and I'm hoping that more of us gain that journalistic like skill set as we use more AI products.

Kaitlin LeMoine: Yeah, that's really fascinating. And I think leads quite naturally in into our The next question we tend to ask on all of our podcasts, given the name of the podcast is workforces, as we reflect on this conversation and see it as you think about your work more generally, what practical steps can our listeners take to become forces and understanding and remaining up to date on the evolution of AI? Right? It seems, as you said, right, nine or 10 months ago, like a shift happened? Well, it's like, very quickly we're seeing continuous shifts. And so what would you recommend?

Siya Raj Purohit: I hear from a lot of people that they feel like they're left behind in the AI push that's happening, like they feel like they don't understand everything that's going on. And I honestly want to say that that's okay, like most of us actually aren't caught up, and that's fine, because AI is not something we need to master as an industry. It's something we need to be able to utilize as like an independent collaborator for your projects. So it's a tool designed for you. It's not something you need to know like all the cool products that are happening and all the thoughts on where the sector can go, like no as an independent user, you should just be able to be like, can I make my work easier? Can I make myself more productive? Can I find more fulfillment in my day to day? Because I have AI as a tool here. And so what I would recommend is starting small, like ask it things about your that are happening in your personal life, like an instructor told me that they had a flat tire, and they're like that led them to the automobile shop that charged them a lot, and they just talked to ChatGPT about all of the maintenance they'd had their car go through. And they asked, What does ChatGPT think? It makes sense, and it turned out like ChatGPT helped them reduce their costs so much by being able to negotiate with the automobile provider about, like, what was wrong. And these are things that, like, some of us never want to get smart in, right? Like, I don't personally care about, like, this space, right? But if it can help me save costs on a day to day basis, on these kind of things, it helps a lot. So think about where in your personal life. Life it can help, and then how it can extend to your work. What are the areas that you don't find maybe fulfilling in your job that AI can help support with the most heart for me, thing I heard was that someone told me who works as a staff person at a university, major university, and they do a lot of like the back end admin work, and they said that meaning in their job is at an all time high now because of AI, because they can push out a lot of that, like admin work they did, and focus on the things that they love, which was like connecting with others, and the events that they host. And so I hope that more of us can find more meaning in our day to day by using AI as an assistant.

Julian Alssid: That's great, Siya. Great advice. It's funny, as you were talking, I mean, one of on the personal front, I've been scanning our refrigerator quite frequently and going AI and saying, Here's what we have. And you know, suggest some effective ways to use use this food before it goes bad. And it's been both both delicious and economical. So I can totally relate. So see. So as we wind down our discussion, this has been just wonderful. Cannot wait to put it out there. How can our listeners learn more and continue to follow your work?

Siya Raj Purohit: I write a lot on LinkedIn about like aI education and the future of work. So that could be a good avenue. And the second is, I run a sub stack called dreams for the future of work, where I'm going to be publishing a lot of the things they're learning from universities as well. And then, of course, if they're interested in ChatGPT Edu, we have a great page and an OpenAI forum where they can learn from professors as well, well.

Kaitlin LeMoine: Thank you so much for taking the time to speak with us today. Siya, it's been such a pleasure, and really appreciate learning from you today.

Siya Raj Purohit: Thanks so much for having me.

Julian Alssid: Thank you. Siya.

Kaitlin LeMoine: That's all we have for you today. Thank you for listening to workforces. We hope that you take away nuggets that you can use in your own work. Thank you to our sponsor, Lumina Foundation. We are also grateful to our wonderful producer, Dustin Ramsdell. You can listen to future episodes at workforces, dot info or on Apple, Amazon and Spotify. Please subscribe, like and share the podcast with your colleagues and friends.