- Speaker #0
Confidence by Dean, season 3. A podcast presented by ISG.
- Speaker #1
Hello everyone, this is Thierry Sebag, the Director General of ISG. But you know, I prefer to say, Dean, we are here on an exceptional podcast. Exceptional because our guest today is an exceptional woman. She is at a distance, she works for a company called Open AI. Chat GPT. Does it ring a bell? Does it ring a bell for all of us? And today's podcast is dedicated to the revolution of artificial intelligence. A revolution that affects education, a revolution that affects universities, a revolution that naturally affects each of us in the exercise of our profession. Will there still be a need for teachers? Will there still be a need for universities? And will our professions still exist tomorrow? Where will they be swept away by the IAEA? To question our guest, exceptional from New York, I have with me two students from the ISG.
- Speaker #2
On my right side, I have Gianmarco. Gianmarco,
- Speaker #1
hello. Hello,
- Speaker #3
how are you?
- Speaker #1
Very good. Gianmarco,
- Speaker #2
where are you from?
- Speaker #3
I'm from Italy.
- Speaker #2
And on my left, Laura. Hello. How are you doing?
- Speaker #4
I'm good, how are you?
- Speaker #2
I'm great. And from New York, we have Sia.
- Speaker #1
Sia,
- Speaker #2
good morning.
- Speaker #0
Good morning.
- Speaker #2
How are you doing?
- Speaker #0
I'm doing great. Thank you so much for having me.
- Speaker #2
Sia, you're working for a company that everybody is so excited to know a little bit more about it. Could you tell us who you are exactly and what you are doing for OpenAI?
- Speaker #0
I work on the education team at OpenAI, and our goal is to bring ChatGPT-EDU to universities and school districts around the world. I've worked in education for the past 12 years. When I was in college, I published a book about America's job skills gap, talking about how American universities weren't teaching the skills that students needed to land a job in industry. And since then, I've been trying to solve that problem, making education more accessible to everyone around the world. So I've worked at different types of early stage startups, invention capital funds, Amazon education, and then joined OpenAI about six months ago.
- Speaker #2
AI is everywhere. And of course, AI is a question for a number of students in the world. Laura?
- Speaker #4
Yes, I have a question for you. A lot of companies have been working on AI for many, many years, but OpenAI has become a leader in the industry within a few years. How do you analyze such a success?
- Speaker #0
I think it's important to note that OpenAI started as a research and development lab, and I think that's still very core to our DNA. Like as a company, we are very academic, very nerdy, and love to kind of think about what the future of the sector could look like and trying to help build towards that. So I think the big success for OpenAI comes from the research personas who deeply try to understand what knowledge work looks like and how we can build solutions to support people who are knowledge workers. And I think the technology advancements that the research team keeps doing keeps OpenAI like interesting and supporting. for the ecosystem. So I think that's a big part of our DNA. And now the business team is trying to make it much more accessible by helping translate some of that technology to how businesses and universities use it. And I spend a lot of time thinking about what the future of universities looks like and how our tools that we're building can help support them.
- Speaker #1
Sia,
- Speaker #2
you mentioned the role of universities and how they handle that new challenge that's all universities in the world have to face. Gianmarco.
- Speaker #3
Yes, exactly. I was wondering as a student, if you can tell me like which is the situation in the United States? Like we know for sure that university decided to use AI and in particular chatGPT. Do you have some feedbacks about this situation like for the first experience?
- Speaker #0
So when I think about AI transformation at universities, I think about it in three stages. The very first stage is the individual stage. This is when individuals became users of ChatGPT and started using it to solve their own problems. So a professor told me that he writes so many letters of recommendation for his students that he created a custom GPT that takes in his past letters of recommendation and a couple of bullet points that are unique about every student. and creates new letters of recommendation. So this sounds like a small use case, but it saves this professor so many hours every month as he creates new ones. So this is at the individual level, people solving their own problems using ChatGPT. The second level is at the team level, so when departments start working together to solve problems that they share. For example, I realized that one university, it needs 40 people need to spend four weeks to decide which class goes into which room on campus. It's a whole assignment that they go through to figure out how to do that matching. And now ChatGPT can do that in a few minutes. And it helps make them much more productive. Like it's a much easier experience for them. So that is at the team level of innovation. Then finally, we're seeing organization level-wide adoption. So this is when universities are like, we're going to have so many different AI touch points across the campus that students are going to engage with. For example, when they come to campus, they're going to have an orientation GPT that helps them navigate the campus, find campus resources, find where the best restaurant is, or how do they change their roommate assignment, being able to talk to the university's knowledge at the orientation level. Then they're going to come to classrooms, which will all have custom GPTs. and professors will upload all of their course material and let students ask the course material different questions. So now in a professor that's teaching a business class, they'll upload their case studies and students can ask questions like, which CEO handled layoffs well? And they'll get the exact examples. So you're basically able to talk to knowledge in a very different way. Then you'll go to career services where career services will like have custom GPTs that let you practice the consulting firms like a recruiter or consulting firms like partner, and be able to take on the different personas to get you that practice before you interview for real. So all of these different touch points will come up and a student will be able to seamlessly move between them, basically absorbing and engaging with the knowledge of the university campus.
- Speaker #2
When we hear what you're saying about how universities and professors use AI, We see a lot of benefits from it and it's true because that's what we expect from it, both as students and as professors as well. I would say that, of course, that questions the role of literacy, Laura, right?
- Speaker #4
Yes, it does. People in Europe fear that students will question the relevance about learning and school in general. What would you tell them?
- Speaker #0
I think that we're just going through a paradigm shift in the same way that we went through when Google was invented and knowledge was organized in a different way. This is the next paradigm shift because now you're able to interact with knowledge and have conversations with knowledge in a very different way. So what my thought process right now is that basically we're going to be able to engage much more deeply and have conversations that students absorb learning in a new way. One really good example from this comes from a professor at the University of Pennsylvania who told me that, Sia, what is the value of an essay? The value of an essay is not necessarily in the output. It's actually in the communication skills and the critical thinking skills that it takes to get to that essay output. So in the past 10 years his class always required an essay submission as a final project, but now he's changed that assignment. that now you have to use ChatGPT to build that final product. But instead of checking the final essay, he checks for how many prompts it takes students to get to an essay that they're satisfied with. Some students are so good at prompt engineering that it takes two or three prompts and they have a really good essay. And some students have to go back and forth 19 or 20 times to get a good essay. And he says that now education is shifting, that that output matters less. But a student's ability to communicate what they're looking for and actually use the tools that are accessible to them, such as AI, is more important for them to be successful in the workplace. And I really like that because I think all of us are going to go through this conversation of what actually matters, what should we teach, and how will our students grow up to become the professionals that the industry needs? And what types of assignments and projects do we provide them to teach them those skills now?
- Speaker #2
Sia, when we hear about your vision on how students use today OpenAI, you mentioned something in ChatGipedi. What you mentioned is something... related to the ability of students to get the relevant prompt and about that new technology, which lead all students to raise a question or to anybody to raise a question. Could you give me your prompt, please? Meaning that in a way, Gianmarco, it's kind of like there was AI skills and specific skills, right?
- Speaker #3
Yes, exactly. So... In order to use AI in a correct way, with AI skills and AI literacy, what would you say to the students or recommend it in order to use it in a proper way? And also, I was thinking to the older generation, what do you think about this?
- Speaker #0
So there's a lot of conversation happening on whether prompt engineering is a skill that we teach or it's something that AI will just get better at, so we don't need to teach it. My opinion is that prompt engineering at the very core is just the ability to clearly state what you're looking to accomplish. And that clarity can come through like upfront work or it can come through back and forth questions that you keep having with the AI until you're satisfied. So a professor told me that he spends one hour prompting ChatJPT to accomplish a nine hour project. So in the past, he would spend nine hours to do this project, which for him was finding the TAM, like the total addressable market for a startup. But now he uses ChatGPT, which he says makes the best TAM calculation, but he needs to put one hour of effort beforehand to provide that information to enable ChatGPT. So this is something that I think all students need to recognize. If you give a generic prompt, you're going to get a very generic answer, and that's not going to be helpful for you. You need to be very specific. So when I talk about projects, I tell ChatGPT like my personality, my role, who I'm presenting to, their personality, what they're looking to get, and all of the additional information I can provide to help ChatGPT do a really good job. And I think that this is a skill that all students are going to need to gain. So just I'll share an exact example, which illustrates this well. Last week, I hosted an event and I asked for audience, like whoever was registering to submit questions for the guests that we were hosting. So the audience submitted. In the past year, I would have read the 120 questions that the audience submitted, and I would have picked and analyzed the best ones and created the talk track. But now what I did was I uploaded that in ChatGPT, and I asked, can you share any of the questions that were coming up more than two or three times? ChatGPT shortlisted that. I'm like, which of these questions do you think are the most compelling? It shortlisted that. I'm like, can you elevate the language around this conversation? Can you frame this conversation in a way that sequentially makes sense? And over time, it basically organized those questions in a way that I was ready to present. And it took me a lot of back and forth. I must have done maybe about 10 prompts back and forth to get to that final output. And it would have been great if I would have done that up front. But it really doesn't matter because you're able to interact and kind of iterate. And ChatGPT doesn't get tired. So I think my advice is for students to just start small. Start. prompting, start asking questions, start thinking about how it can become a knowledge partner for you, and then start building up to bigger and bigger projects that you do. And also use it as companion to think about your learning objectives and long-term goals. One of the best conversations I have with ChatGPT, which I recommend to everybody, is tell ChatGPT what your goals are. Maybe it's a goal that you have at a five-year mark, like in five years I want to have XYZ position at this type of firm. Share your current resume or LinkedIn profile or even just bullet points about who you are right now and ask ChatGPT to create a one to two year roadmap for you on how to accomplish that goal. And you'll be amazed by the types of answers it gets. So these are just some of the examples of how I would recommend students start using it immediately. And then same for the older generation. Like I have heard so many unique examples of how people are using it in their lives to answer questions about like. their house or to answer questions about like the shopping lists or travel plans. Just start incorporating it and you'll see a lot of benefits on how ChatGPT can help make you a better professional and help support your personal life as well.
- Speaker #2
When we hear what you're saying, Sia, we can perceive the whole benefits of AI and of chat GPT for everybody in their daily life. Of course, that requires skills and ability, as you said, for making a class of nine hours, it might require an hour prompting. And of course, the more you will prompt, the better you will be at prompting, of course. just like in any technology. But we are just at the beginning of this revolution that we perceive for the future. How do you see the future for OpenAI and what are the challenges ahead?
- Speaker #0
In the education space for OpenAI, I am really excited about us continuing to work with universities, schools, districts. I think we can do a lot to support like presidents and provosts who define visions for universities, professors who help communicate the best learnings to students, and then also students as they go through their academic journey but also their professional journey in the future. So really excited about continuing to build tools that will help enable those AI touch points across campuses but also help students absorb that learning in new ways and hopefully very excited about this multi-modality piece as well. Soon we'll have like video and audio and images and text all combined together to help students learn in the way that's best for them. If they learn best through like podcasts, we can do that. If they learn best from written text, we can do that and basically adapt to each student's needs. I worked in education for 12 years and personalized learning was always like the holy grail. Like that was what we always wanted to accomplish as the dream. And now we have like I have a personalized tutor that I talk to every day. It knows of my projects, it knows my aspirations, and I think our goal is to help bring that personalized tutor to everyone around the world. As far as the challenges go for OpenAI and AI in general, I think one of the big things I think about is like adoption of AI and making that equitable. So a consulting group published a case study about how their lowest performing consultants started performing at the same level as their highest performing consultants when given AI tools. And that just showcases the gap. So for students who use AI, we'll be far ahead of students who don't. And this is true for all knowledge workers as well. And so I think a lot about AI literacy and helping students just get started with using this. So then they can be competitive in the workforce and hopefully achieve their potential.
- Speaker #2
We are the end of this podcast dedicated to the revolution of AI with an exceptional guest. from New York working for OpenAI. Her name is Sia Raj. Thank you so much for all these insights you shared with us about the future. A future where of course AI is the key of personalization of education, the key for giving education to everybody, the key for providing the best education, the best classes in the world for providing the best to all of us. Of course, that's a help for everybody. Of course, it might be a danger if we do not take into account the risks and the danger of AI. We should be aware that when we use AI, of course, there may be deepfakes, there may be plagiarism, there may be cheating, of course, but that's... new technologies that today all universities in the world are using. They are using it just because we know that in the future everybody will use AI. That's something that will be embedded and just like we're using Microsoft Pack Office in our daily life, tomorrow we will use AI on a daily basis. That's probably a message of hope for everybody.
- Speaker #1
See ya!
- Speaker #2
Thank you so much for all of these insights.
- Speaker #1
That was Thierry Sebag for this exceptional DIN Confidence with a brand newbie from New York. Her name is Sia Raj. We asked all our questions. Do you want to know more about this IA revolution and how at ISG we integrate artificial intelligence into education, personalize programs? Easy! On vous dit, retrouvez-nous sur LinkedIn. Nous, on vous dit à très bientôt. Bye bye.
- Speaker #0
A podcast presented by ISG.
