E27: AI at WPI | Elke Rundensteiner and Yunus Telliel

Colleen:  A-I, they may be the two most popular valves in technology today. You can't open a news site, attend a conference, or sit in a classroom without hearing about artificial intelligence and how it's reshaping nearly every part of our world. But behind every AI system, every algorithm, every model, every decision is a human being and the world needs professionals who understand not only how to build these technologies, but how to build them responsibly. In this episode, we take you inside the design of WPIs new AI degree, why it's being launched now, what makes it unique, how it prepares students for a rapidly evolving job market and how ethics is woven into every part of the curriculum. Let's get started. Professor Elke Rundensteiner is one of the key architects behind this new AI degree, and she's been deeply involved in designing a curriculum that blends strong technical foundations with the hands-on project-based learning. WPI is known for. Elke, welcome. First question. There's a lot of conversation about the demand for AI professionals from your perspective. What makes this the right moment for WPI to expand its offerings in artificial intelligence? 

Elke: So there is a desperate need for AI talent. And why is it right for WPI? Because we have been getting ready for that moment, I would say for decades. Uh, because we have the right faculty in place, we have the right classes in place, and we have the right projects in place. 

Colleen: How do you see AI transforming industries and. How is higher education adapting to keep up with that pace of change? 

Elke: Many, many aspects are changing. Any company that has digital data of some form is now able to leverage that data in much better ways, be it for discovery, for better product management, being for marketing, be it for better services for average citizens, as travel services, for example. So all industries are being transformed. We are really at a disruption here, so that's a very exciting thing. How do we keep pace? It's challenging, I have to admit it is not easy, but our faculty are ready for this challenge. Especially in computer science and in ai, we can continuously changing over and over when you think back over the last 20, 30 years. And so our faculty are very agile. They learn new things. A second way that we are keeping pace is we work closely with industry. We have very close industry relationships through an advisory board, through projects where our students work with industry in a collaborative fashion. So we see as the changes are happening, we understand them and we bring them back to our classroom.

Colleen: You built up a very influential and important advisory board for data science. Is the same thing happening in ai? 

Elke: Yes and no. This advisory board, data science has been expanded. Because many of them actually serve dual roles of data science and ai. They're very close sisters and brothers of fields. So we've expanded the members on that board, but it's the same board that's now serving data science and ai.

Colleen: That's great. They're evolving with how they're moving into AI and then feeding that information back to you and the other faculty members. That's correct. Many of their jobs have equally changed because they are in the middle of these industries with major changes. That's important for the students too, to see how these people in industry have to shift what they're doing, their understanding, their skillset, much like our students are now doing.

Elke: It's exactly the same. The students that we are producing need to be lifelong learners. They need to go out there and work in that industry, and that industry is gonna move on them by the months, by the year. Like what we, what we are teaching them now is really how to learn and how to love learning. And the advisory board are in some sense our past students that are now successful. So they're all continuously adapting. They're exciting role models in that sense. Yeah, and so I, I make sure, and there's advisory board meetings that some of our students are participating and get to interact with the board members and learn from them as well.

Colleen: Good. And not just taking that information, probably pushing back and asking questions too. Asking questions, yeah. Yeah. On that advisory board, there are some well-known companies represented, correct. 

Elke: Yes, that is correct. On our advisory board, we work, for example, with Microsoft. We have had an IBM representative. We work with a number of insurance companies. For example, MassMutual. We work with MathWorks. They have a cutting edge machine learning tool available, and a number of startups as well. 

Colleen: Great. Some big names there. So what gap does this new degree fill that the existing programs WPI has in data science and computer science don’t fill?

Elke:  It just offers one more opportunity and pathway. So for this degree, unlike let's say in computer science, they can narrow down and learn the skills specific to ai. For example, they will learn the computational skills, they, they've learned the mathematical foundation skills that ping them together and then learn AI. On top of that, they learn the machine learning foundations, they learn the machine learning methods. Then they learn advanced techniques and they led some focus more in depth on the AI field, other degrees, for example, data science, um, has also machine learning techniques, but they apply to business was really three fields coming together: Business, Computer Science and Mathematics Foundation. And here we are really focusing in this new degree. We are really focus on the AI fundamentals and the cutting edge AI tools. With the foundation that AI is built on our students learning will be built on. Yeah. Yes. 

Colleen: Okay, so what makes i's approach to teaching artificial intelligence distinctive from other universities that are launching similar programs. They're popping up at a lot of places. 

Elke: WPI is very special. We have complete project based education, and of course for this AI degree, project education is at the core of how we approach the education. So from the very beginning on, they don't just learn the theory, they apply this in a variety of projects in classroom continuously. That's really what makes it unique at WPI. Secondly, we have cutting edge researchers that are working in ai. We have experts in large language models, experts in deep learning, in reinforcement learning, and all the cutting edge areas. And these researchers are bringing back their latest ideas and the latest knowledge back that I came to the classroom and, uh, working closely with these students. All these students will do a project where they're actually directly taking the skillset they just learned and they're putting them to practice. And that is what makes it very unique, because by the time that they're done and they're graduating, they hit the ground running, they really know the material, but they know how to apply it.

Colleen: Many of the listeners of this podcast understand MQP, our major qualifying project. Can you explain a little bit more about the importance of having our students do an MQP in AI to see it in real time? 

Elke: We have remote project sites, for example, in Silicon Valley, for example, in Switzerland, for example, and Wall Street. So we have remote project centers where our students get embedded into real world projects. They work on a project that matters to a sponsor. It can be several BS in AI students, but they can also work in teams where there might be one AI engineer, one AI student, and one computer scientist, and one data scientist working to together on solving a problem. It is critical for the education, I do believe, because they understand the real world problem that a company cares about, and then they gain self-confidence because they learn how to take those skills that they know and really put them into practice. Um, that's the best way to learn to learn by doing.

Colleen: Absolutely. So how do you and your colleagues collaborate to design such an interdisciplinary program? 

Elke: We are really collaborative at WPI, and I would say that I strongly think this degree is one of a kind and has a collaboration across 12 different units at API and all schools. We succeeded to achieve this by the data science and AI faculty putting together a framework where for this degree we have core courses that teach them the fundamentals and the technical skills to learn about AI tools, technologies, and to apply those technologies. Uh, but then we left enough flexibility in this curriculum so that the students can go after their passion. So if a student wants to dig deeper and deeper into those tools and maybe go to grad school in ai, we have a pathway for them. They can take more advanced classes and they can even take our grad classes that are cutting edge. If a student wants to take those skills and apply them to another field of their passion, this could be in social science-making a difference on society, it could be in engineering, it could be in FinTech, whatever it is that they want to solve. We also give them pathways in this degree, so we provide enough flexibility that they can learn in second field and application field and learn how to speak with those in that field, to understand the terminology, to understand what problems is that field trying to solve, that later they can engage and take a job in that intersection where there still would be an AI engineer. But they would collaborate in that particular application domain, be it a FinTech company with experts in that field to solve AI problems for them. So this is very, very unique. I don't think that there are many such programs. In the world. 

Colleen: It is a very technical program, but the fact that the students, before they graduate with their bachelor's of science degree, they're already applying that to what they're passionate about. As you said, with social science and policy studies, they can be looking at. Poverty levels, homelessness, hunger with an AI lens or FinTech, does that expand in your opinion? The type of students that might be interested in ai. I believe it will draw a large variety of students.

Elke: It will draw students that are deeply interested in the technology itself and want to be an AI engineer and be the very best AI engineer that can, and maybe even advance these tools in the future. But I think it will also draw students that are interested in AI as a technique because they want to use it. To do something else. So it can also be drawing applied AI students that are actually going to learn sufficient amount of that technical skill and then can collaborate with others to solve a problem in a different domain.

Colleen: That's truly wonderful. So it's not just all about zeros and ones or bits and bytes. Yeah, it really gives breadth. It allows both the core AI skills and also applied AI skills. The graduate program that was launched two years ago has seen some overwhelming success in the amount of students that are applying and have become part of the program. What is some feedback you've received from those that are going through the Master's program or even from the faculty members? 

Elke: I would say the students are voting with their feet, whatever the cutting edge classes are that we have added. For example, we added an ML ops machine learning operations class.

We added a generative AI class. These classes have waiting rooms to get in. Students are really interested to learn the latest technologies, and so it's highly popular. And it made perfect sense to expand this to the undergraduate. So it makes, yeah, so it gets us ready because we now have offered these really advanced classes that we put together two years ago, and now we can take some of those and move them down to the undergrad level.

Colleen: I'm going to go back to the fact that AI is advancing so rapidly, even what we knew six months ago. Is changing or has changed. So how does WPI stay on top of these developments in technology?

Elke: I would say it is in the DNA of a computer scientist and the noun and AI researcher and faculty that change is part of the normal game. Things are continuously changing and that's what keeps our job fun and exciting, right? So we want to learn new things. So we all faculty are curious. They want to understand the latest technologies and techniques and then share them with students. That's what makes it fun. We don't want to be teaching something that's in a textbook a hundred years ago. Repeat that over and over. So faculty are learning new things. They conduct research, they work with industry. They learn about the latest and they bring it back. That's what's so exciting about being in academia. 

Colleen: And I'm guessing you, as you're helping to design all of these programs and curriculums have learned a lot too, do you feel in some sense that you've had to go back to school to learn how to best offer these programs?

Elke:  I think that there are two sides to this one, just the research that we do and the project experiences we do with students. Every time we learn something new, we learn something from the students. As they dig out more ideas, we learn something from industry partners. So it's been a continuous learning experience and we also learning how to deliver this material better. We are learning that students want to be engaged much more. They want hands-on projects. They don't want to passively sit in a classroom. 

Colleen: It keeps us young. What excites you the most about the potential impact of this program for the students, for employers and for society? 

Elke: AI is changing the world. We're really in a disruptive moment, and I'm really excited about the fact that we can be training the next generation that are going to be the change makers. They're going to work on whatever comes next, and I am confident that they will make these changes for the better, for the better of our jobs, for the better of our society.

Colleen: Final question. Maybe you already did it with that answer, but if you could summarize the heart of this new degree in an elevator pitch. What would that be? The degree is unique in that it's both rigorous and flexible. It gives the students the underlying theory and techniques that they need, but it lets them chart their own pathway. Each student, I am convinced, will take a different path. They will all find their passion, but they will have the rigor behind them to succeed and to continue to learn. And we are setting these students up for success. What could be more exciting than that? 

Colleen: Thank you so much for your time today and really getting to the heart of why this is so important, not only at WPI but society in general.It's been a pleasure. 

Elke: It's been a pleasure. Likewise. Thank you for having me here. 

Colleen: Now that we've heard how the new AI program was built, we want to shift into another part of the degree that makes it truly distinctive. As we've seen in industries from healthcare, to finance to manufacturing, AI's influence is growing fast, but so are the questions around fairness, transparency, accountability, and impact. WPI wanted a program where students learned to build powerful technologies. And understand the broader systems they influence. That's where our next conversation begins. I'm excited to welcome Yunus Telliel to talk about what it means to teach ethical thinking alongside cutting edge AI. Yunus, welcome. I'm just going to have you introduce yourself, your department, your areas of study.

Yunus: That's actually the most difficult question. I didn't start with a softball. I'm an anthropologist and I'm not in an anthropology department. I'm in the humanities and arts department, and I have affiliations with interactive media and game development program and robotics engineering department, and I'm saying difficult because I like this difficulty of describing myself because it really helps me highlight the complexity of the work I'm doing and all the interdisciplinary connections. So, um, so I'm an interdisciplinary scholar at WPI and I think WPI is a perfect place to be an interdisciplinary scholar, especially thinking about AI today. 

Colleen: Yeah, I was going to say that if there is a time and place for someone interdisciplinary, it's WPI when it comes to AI So. My first actual question is what is ethics in AI? 

Yunus: I think that is the million dollar question, perhaps, because that's the most important question, but it's also really difficult to wrap our minds around it. One of the, I think, difficulties is the scaling. You can think about ethics from the perspective of the user. Me turning on, let's say chat, GPT, and entering a prompt and thinking about what comes out of it. You can ask a lot of ethical questions about the accuracy of the information and the kind of maybe cheating. That's obviously a concern in the US and also globally right now. And that's one scale. And then if you actually zoom out and then think about the internet networks and the data infrastructure, data storage. Now you also talk about physical spaces, environmental impact, and so that's a different scale. And then if you think about the chips that actually go into these physical spaces that are used in the data storage places, then you actually talk about extraction and mining. So ethics in that way. It's perhaps the most important question about the biggest challenge for ai, but it's also really one of the most difficult questions for us to tackle, perhaps because of this multi-scale orientation of the problem.

Colleen: Yeah, it's like that game, whack-a-mole, just when you have one figured out, something else pops up. Exactly, yes. When did you start studying ethics?

Yunus:  I have always been interested in politics and ethics of science, especially as it relates to religious communities. I was interested in the perception of science and that obviously you might remember COVID-19 vaccine debates, and you see a lot of ethical concerns about public health that are framed in religious like idioms. So that was the kind of my interest in ethics and science. But my main interest in technology, AI, robotics started when I arrived at WPI. I met a lot of really bright folks, graduate students, working on perhaps still technology of the future. And they kept asking me questions about the ethics, social implications. So in life we make a lot of choices as we go through certain difficult situations. And then we do perhaps a cost benefit analysis to navigate between different costs, this damage, maybe that damage which one is more reasonable, right? Yeah. And my problem was why don't we actually think about the design and development of AI technologies and robotic technologies so we can actually make things designed not to harm people. So bringing ethics as a design principle, building design in the system rather than as an afterthought, rather than as an analysis of something that has something bad that has already happened. Why not actually think about the ethics? Before you even build, if you think about ethics in design, in the very development in the lab. So that to me was my main, I think the push point. That's how I became a scholar of AI ethics. Mm-hmm. And, uh, the work I do is shaped by that interest and in how we do things differently in lab, in research and development. So that, that will be products out there. Technologies out there, services out there that do not do damage.

Colleen: Got it. Has it been frustrating, challenging, maybe even a little exciting for you as someone that looks at this technology with ethics to see how quickly AI has taken over almost every part of society and. There's a lot of catching up to do when you're thinking about the larger picture, as you just mentioned, and the impacts may be 5, 10, 20 years down the road, if not more.

Yunus: That's correct. When we think about ai, we have to think about the future because. This is one of those technologies that has immense impact. If you get something wrong, if you have a problem in what you have, it could impact millions and billions of people. So we are talking about technologies that are really prevalent in our societies. So for that respect, any mistake that is in the design or any unethical issue. In the design of these technologies would change the lives of so many people. And so that to me is the ethics question, has the potential to be scaled up so immensely and AI in that way. We have to think more carefully about AI systems for that reason, because we now face a technology that could impact lives of many others so quickly and so fast. So in that way whatever we do in the present is linked to what happens to people in the future, and I think that timeframe or the time between the future and the present, uh, for these technologies is very short. 

Colleen: What has surprised you? Are students thinking about how AI can be applied, can harm, or make things easier?

Yunus: I think there are so many different ways in which these ethical concerns play out in student projects. So sometimes students are so excited about using AI technologies to solve a problem. So they say, okay, if I actually collect. Some data on some information that I can use for a nonprofit helping immigrants, but when they build an application for immigrants in a situation where those immigrants’ identities might be protected, might need to be protected. Then they have to navigate between the concern of building a system that is operating effectively. Maybe with more data, with more information put into it. And then the other concern being the privacy of these individuals. So when you have to navigate between these two conflicting concerns, two conflicting values, perhaps effectiveness of a system and privacy of individuals. How are you going to navigate this? So a lot of student projects that I advise and in my classes, I highlight that there are no easy solutions to these projects. And I think WP is a good place for students to understand that these kind of problems coming from real context. I think the project based learning is definitely a good place for students to experience this. As they do the projects, they can see these different conflicting ethical issues emerging and learn how to navigate them, learn to weigh them against each other, and then they need to make a decision about the technology they are building. So maybe they don't need certain types of information. Right, because the data will be individual privacy issues.

Colleen: What about jobs? There's a lot of threats to the future of jobs, job creation in the age of ai, but on the other side of the coin, the loss of many jobs getting taken over by ai. How do you work with both students and faculty to keep that in mind? 

Yunus: That's a great point. What is happening with AI definitely highlights the value of social skills, understanding of communication, understanding of context, and in that way, I think with WPI general educational philosophy. there is so much emphasis on problem solving. Within specific contexts, and I think that really helps students build certain types of skills that adds onto their technical understanding of AI systems. And then on top of that, they can actually manage expectations with respect to. How to use a system for a particular purpose, because then now you can, they can bring the human factor, they can bring an understanding of the social context, they can bring an understanding of the ethical consequences of the system. So that's a layer of skills that are wrapping around their technical skill. To me, that actually. Is distinguishing. So you are not just a technical expert, but you are an expert in the broader sense of the world. Because we are dealing with a technology that is not just technical anymore, it is social, it's cultural, affecting politics. And that way we have to expand our understanding of expertise. And that's what I see with our students. And I think my work and my colleagues work with students. Often focus on that holistic perspective in which the students have the excellence in building and using AI systems. That involves not only that they are really good at using or building AI systems, but also they know when not to use it. They know the limits, they understand opportunities with respect to the limits or the consequences. So to me. This is what WPI has done really well. 

Colleen: A lot of faculty here are very much committed to that holistic individual, holistic person that is at the center of STEM education in building a curriculum for ai, whatever shape or form, not only at WPI, but at other colleges and universities that might be looking into this. Should there be one course on ethics and AI? Or is this something that has to be a part of every single curriculum when you're looking at building technology or understanding technology? 

Yunus: I think that's really important. What I said about scaling, how AI technology scale our ethical understanding or the questions also scale and with that respect, every department, every program with respect to the work they do with AI, and that could really differ from each other. How biomedical engineers think about AI wouldn't be the same, you know, as data scientists. I think of AI as a system of systems because there are systems that we deal with in our work, but there are also larger systems in which those individual systems operate in. So for that respect, I believe that all of us should have something valuable to bring to the table because this AI systems, AI technology right now manifest itself in many different ways, depending on our needs, depending on our context. So in that way. Ethics question should be addressing those specific manifestations in engineering, sciences, humanities, and arts. But I think what is really valuable though, not only just addressing ethical issues with individual courses in different disciplines, but how do we bring all these different concerns to each other to speak to each other? And I think that is the ideal thing. It would be the platform. Of ethics modules or courses or work done on ethics, and they come together and they talk to each other. They enhance each other's perspectives, and I think that this is already happening at WPI. We need to do more work to build this platform to learn from each other. For instance, a lot of my knowledge in AI ethics comes from folks who are in technical disciplines. I ask questions and they help me understand. A technical issue, but I can't see how that relates to social issues. That's not always possible if you are an anthropologist like me, right? In a different institutional setting where you don't actually interact with technologists, engineers, or scientists. I think WPI has that potential to highlight this conversation that can happen on ethics. So I teach on ethics. I know that my colleagues offer ethics modules in different disciplines. I think while we are trying to build a conversation in which we can learn from each other, we can't be siloed. 

Colleen: I really like that because I came into this conversation thinking that it should be these technologists keeping ethics in mind as again, they're developing, they're producing, they're adding to products, technologies, approaches, but you have to talk to one another, right? Because you can't be explaining ethics of a technology or a future technology to them without understanding what they're trying to accomplish. Thank you for sharing that. As we're wrapping up, what would be your advice to a student? In high school, maybe even middle school, who's looking at STEM fields and wants to make sure that they're doing that responsibly. 

Yunus: I actually receive emails from high school students and they often have great CVs already. They have been in the robotics clubs, they've been working on AI, they're really good, they're ready for college. They have the skills, but they're also really worried about the impact or the consequences of what they're doing. So they tell me, Hey, professor, I'm interested in robotics. Can I come to your class or can we just chat because I'm working on this project, but I don't know the consequences of what I'm doing. That to me is amazing because they do not only see AI as a tool, as a technical system, but they see it as a question. And that to me is the beginning of any ethical inquiry. Not to settle on something that is given to you, a tool, a technical system, but to turn it into a question. And I think in humanities and arts department in global and integrative studies, in social science and policy studies in in interactive media and game development, we have actually a lot of people thinking about AI in what you might call non-technical areas. Mm-hmm. And that's actually helping us to turn AI into a question rather than a something that is already done that's already finished. So I think keep asking questions is one thing that high school students can do. 

Colleen: Okay. I'd like to end there on a high note that as scary as this technology could be. It sounds like people are asking the right questions. Yes, hopefully many of them. Alright. That's right. Thank you. Thank you so much for your time.

I really appreciate this conversation. Thank you. 

Thanks for joining us for another edition of the WPI podcast. We hope you enjoyed this discussion on AI at wpi. To learn more about the new bachelor’s program as well as the masters and PhD offerings, visit wpi.edu and search for ai. There's a lot of good information available. To hear more WPI podcasts, visit wpi.edu/listen. Or search for the WPI podcast on Spotify, apple, YouTube, or wherever you get your podcasts. A special thanks to our guests, Elke Rundenteiner and Yunus Telliel, as well as our audio engineer Varun Bhat. I'm Colleen Wam