Episode 2: Yu Chen

[Music]

Vincent del Casino: Hi there, my name is Vincent Del Casino and I want to welcome you to the Accidental Geographer. On today's episode, I'll be talking to Dr. Yu Chen, who is an associate professor of information systems and technology in the Lucas College and Graduate School of Business. This is going to be a really interesting conversation as we cover a wide range of topics around human computer interaction, artificial intelligence, and the questions of the social good. And how all that plays out in the classrooms here at San Jose State University and across the California State University system. So come on board. This is gonna be one awesome discussion. All right, well, let's jump on in. Thanks for being here. I really appreciate it.

Yu Chen: Thank you for having me here.

Vincent del Casino: So I'm excited to have you in this conversation with me because you're doing a number of fascinating projects coming out of the School of Business, but thinking about questions of the social good. I wanna kind of start though with a little bit of how you got interested in the kinds of things you've got interested, and your own academic journey, right? From China to Europe to here. And just get a sense of, you know, what got you excited about the kinds of topics you've gotten excited in and how you've kind of evolved over time in relationship.

Yu Chen: Yeah sure, maybe just a little bit of background of my academic journey. So basically I had my, I did my undergrad in information security in China and that time I got interested in my dad's Nokia N73 phone and when I talk about that phone to students they never heard about because it's a little old, but that got me fascinated because This is the first time I saw a phone that can not only take phone calls, but also take photos and play music. I know it's before iPhone. And that's made me fascinated about, oh, what is the hometown of this brand? And then I got to Finland and Norway for my master's studies. And during my internship, while doing the master's degree, I was working actually at Nokia, where I got know, oh. Technology is not just about technical components. There's a huge human factors in that. And that's motivated me to do my PhD in human computer direction. When I get to know how to design anything, not only technology, but also curriculum and everything with user centered. So that's where I have a kind of design thinking or user experience mind in mind before I joined San Jose State. Well, I spent a couple of years at UC Irvine continuing doing research. And all those bit HCI, AI, etc. All right. That's before I came to San Jose State. I have to say, there is so much. I feel grateful for the resources at San Jose States. The journey I got interested in AI was actually the second month when I was here, the third month, maybe the first semester. I was introduced by one of our colleagues, Larry G. Who introduced me to the IBM call for quotes. And this is an international competition that's designed to encourage coders, programmers worldwide to use IBM technology to create prototypes for social purposes. And I was there for the final award ceremony where we see how developers from all over the world are creating system for hurricane, for earthquakes. Wildfire and I just feel this is so cool. We can do that at SJSU as well. So that's the start of the story.

Vincent del Casino: That's awesome. So when you think, I mean, I think human computer interaction is so important because we're extending ourselves every day in relation to them. When you first got into that, what were kind of your foci? What were you looking at? What questions were you asking around computer, human computer interactions?

Yu Chen: Yeah, the first thing I was always curious is, because I came from a computer science background from undergraduates, and the main focus is how we can develop a system that's useful from a developer's side. And then once I did this internship at Nokia, I realized users might not care that much about what functionality you have, but more of what you can offer to them, what they can benefit from this service or offering or technology. So the first question is, Can we design with users in mind? And that extends later to when I started teaching at San Jose State, how can I design curriculum with students in mind, especially who the students are, their background, their cultural background, and their diversity in their discipline, et cetera.

Vincent del Casino: So if you think about that question and start to push on it, so computer scientists can build all sorts of things, but whether or not they're actually gonna be used is an interesting question, might not always be at the heart of that initial computer science question. So this field of human-computer interaction has kind of evolved over time to ask these bigger usability questions, right?

Yu Chen: Exactly and then later extend to more of ethical and technology ethics all involved and later and also on the media we heard a lot about AI ethics etc. It's all along when we're keeping users in mind number one can we offer them the best benefits and number two how can we make sure that we are not putting them at any kind of risks and including privacy and bias and etc.

Vincent del Casino: What's really interesting that you landed in a business college, right, because you probably could have landed in a lot of different places. Could have landed in an information school, you could have ended in computer science, you could've even landed, I think, in computer engineering. So what attracted you to a role in a management information systems program, you know, in a Business College, as opposed to some of the other places?

Yu Chen: Yeah, sure. So actually, when I was on the job market, I literally interviewed at computer science, department of art and design, and also the health informatics and including and then also business school. And to be honest, I think the most important reason is the collegial atmosphere I feel at our department. And that's the major, major reason I have to say. And then on the on the top of that. When I was doing the demo teaching in the classroom, I can literally feel the different kind of spirit from students in terms of their innovation and their big picture thinking. And I was feeling, yes, that's something I want to, it's sort of nurturing me as well. So those are the major two reasons I ended up choosing San Jose State, especially School of Information System Technology at College of Business.

Vincent del Casino: Well, it's really, that's really fantastic. And I know that accreditors in business, they actually look at what are the social impact of the programs and things like that. So I imagine that was also a driving force in this. How, given your interest in human computer interaction, its usability, its value, the ethical questions, that social good comes out of that question, does it not?

Yu Chen: Yeah, definitely. And that's exactly for business students. I think there are two components that are keeping mind in the teaching component. Number one is I'm hoping to teach them or empower them to learn some of the technology which are needed, especially in the Silicon Valley. And the other component is business students, they are trained to be the leaders and innovators for the business and the society. So. That's where I started to think, oh, this is a really good marriage of how we can design something that can marry this together.

Vincent del Casino: Yeah, that's really great. When you think about, I want to shift gears a little bit. When you about artificial intelligence. So I've been writing about this and I've talking about it. So I don't think it's either artificial or intelligent yet but we use the word a lot. To you, what do you think it really means? And where are we in the context of artificial intelligence?

Yu Chen: Yeah, so for me, artificial intelligence is emulating human senses and also the cognitive processes. And for example, we're thinking about human senses, human can see and humans can hear. And before computer, before artificial intelligence, the computers will not be able to see a picture because for computers, the picture is just a bunch of pixels. And for the audio is just a lot of digital processing and digits there. So it's not until AI, now what's hot on the news is the computer vision and natural language processing. And those are empowering the computers to be able to sense and also make cognitive decisions. So for me, it's more of mimicking human intelligence in terms of sensing and interacting with the world and also making decisions. In terms of where we are, in general, in... AI, they're putting them into broadly two, three categories, and sometimes maybe two, depending on which company are making the definition. One is narrow AI. And for narrow AI, we have already been interacting with that for the past decades. If we're using YouTube to get recommendations from videos and Spotify for music, Netflix for movies, we're already exposed to recommendation, with recommendation algorithm, which is the narrow AI, which is part one. And part two is more of broad AI, where right now we see more in the, for example, self-driving cars, and also right now, we see the chat GPT, et cetera, which seems to have a little bit more advanced capability. And the third category is AGI, artificial general intelligence, or general artificial intelligence, and people sometimes use it differently. And that's where we believe AI is able to actually... Make all decisions at the similar level of human intelligence level, etc. But that's according to some of the AI experts in the domain. It's not coming until 2050. Again, that's the prediction. It might be sooner, it might be later. Right now, we're in the second category of broad AI. Where was the AI used in more and more domains capable of doing more tasks?

Vincent del Casino: I think it's important, and I think what's a little bit lost in the public conversation is that nuances, because we're still, I mean, I'm asking this as a question, I think we're still in a place of really deep machine learning as much as we are, right? As you said, general AI is partially about computing power and whether or not we're going to build different kinds of networks than the classic ones we have to be able to do that sort of thing. Is that a good way to think about that?

Yu Chen: In terms of the direction where we are.

Vincent del Casino: Well, and even how we teach about it, because I think we use the terminology. And I imagine when you teach, you differentiate these types of, quote unquote, artificial from narrow to broad to general, right?

Yu Chen: Mm-hmm, yeah. And I think there, usually when we're trying to give students this kind of picture, we're talking about, hmm, actually artificial intelligence started, well, some people would define it started from 1900, when cognitive science is coming to place, when people understand a little bit, and then there's a bumpy road and gradually developing in the last, actually, 100 years. Now, the question we're always asking is, why now? Why now it's developing so fast. And usually there are three, there are different components. Number one is the huge increase of data, the amount of data created from the sensors, from social media, et cetera. And the other is the computational power, right now we're see dramatically improving. And the third one is, the algorithms because of the computational power and the data. And we are able to have more advanced algorithm, and right now people are talking about larger language models, etc. So those conditions are not mature until the past few years where we see the rise of AI.

Vincent del Casino: Yeah, it's really interesting. So I see that you've become really interested then as well in pedagogical approaches in relation to artificial intelligence. And that's really driven a lot of your work, including your published work. What got you excited about that? And how have you thought about how you're bringing these questions into your classes? And then academically You know, and intellectually, how are you writing about them? What are you talking about?

Yu Chen: Yeah, sure. So essentially, I feel the integral part of the teaching service and research here at San Antonio State. So I started, mentioned earlier, I attended the IBM call for code and was just thinking, this is so cool and we can do something among our students as well. So gradually over the past four years, we're gradually implementing by bringing AI At that time, back in 2018, not many students, actually business students with all diverse accounting and management, finance background, they might not always know about AI much. So we have to think about what might be the content that we can deliver for students who are not most well-known for technical background. So that's the first component. And the second is we definitely see this kind creation, creativity and innovation spirit from students. How can we include that as well? And the next component is, how can we incorporate the learning that's not just a one size fit all, but they actually feel motivated. And that part, we're able to tap into their own culture and community assets. So when others are coming together, I feel it's tying back to HCI, we need to understand students' body here. And with that, we just do one experiment, maybe in fall 2018, and then say, it's really well received by students. Let's try again in 2018. And then later, we bring experts from the industry, from IBM, from Google, and from Microsoft, for them to see, look at ourselves as these students work. And then they're impressed. So I feel, OK, it probably is working. So we continue that... Until COVID hit with the field, maybe it's not no longer working, but actually it's working way better than before COVID because there's so many social problems during COVID. And that's went on until 2021. I get a chance to meet some of our collaborators that's currently on our NSF grant, including Dr. Frank Gomez at the Chancellor's office. And then we have four excellent CSU. For other CSU faculties. And we came together to see how can we take this to benefit more students in California in CSU system. So we came up with this idea of maybe proposing a larger project. So that's kind of the journey how it started from the very beginning was just the idea and then gradually grow and grow until now. And then back to the question about publication. Actually, at College of Business, we really support pedagogical research and educational research in general. So as for me, because I came from an HCI background, and my prior focus of research is more of designing systems and we evaluate with users and then see what we find and then just to see, number one, is that working or not? Number two, how and why it is working. So for the pedagogy, it's exactly the same. Here's the design curriculum and we. Experiment with students and then see and collect some data to see whether, how and why it is working. So this is exactly the same philosophy that's powered my research work as well.

Vincent del Casino: That's so interesting. And it's not common to see a business professor on a National Science Foundation grant. I think it's fantastic, but it's so exciting that you're bringing that component in the way in which you're thinking about things and how the value of that research enterprise plays in. I wanna pull on one of the threads that you hit, because in your paper on the innovation farm and about teaching artificial intelligence, you talk about gamified social entrepreneurship. And you mentioned that getting students to see themselves get excited. How is gamification played into this? And how do you see it? And what is its pedagogical value in the classroom?

Yu Chen: Yeah, definitely. For our business students, I can definitely feel from the very beginning, they have this kind of nature to be the future leader and innovators, et cetera. And then we really want them to feel this kind of how that might feel like. So just give a little bit background about the innovation farm, which is a former name of AI for social good. So those are some of the prompts for students. We teach them the basic AI. Concepts through demos and applications and also use cases. So this is your toolkit. Imagine that they know what these technologies are and what they can use for them. Imagine they're the business manager and the project manager in the future. Then later, invite them to think about what might be some of the important social problems in your community, and then they need to come up with solution. Before they start to present, They also need to know what is the role among the team members, the working teams. And then they will come up with their quote unquote identity. I'm the CEO, I'm CEO, and I mean, so they feel motivated because they feel the sense, oh, I am a leader and I can imagine what the leadership might look like. And they create their logos for their quote, unquote social enterprise or quote unquote social startups. And then, they actually, as we mentioned, we invite the industry professionals to come judge their work, and they are the quote unquote investors. And the winners of the final competition, they get checks, quote unquote. And the checks, formerly, they are checks from teacher's bank, where we have extra credits and things like that. And later the checks actually become actual, basically, the students are able to have the money, but not for themselves, but they have the. Freedom to choose one nonprofit organization that they feel align with their motivation and intention so that we can donate that amount of money to the nonprofit organization that mesh with their interests. So I just feel want to provide as much real world experience to the students through this kind of gamification.

Vincent del Casino: That is really interesting. So, and I love it because one of the biggest challenges in retention is how to get students to see their own ownership, right, of their education and their experience. And I have a colleague at Michigan, Barry Fishman. I don't know if you've ever seen his work. He works at the intersection of information science and education. And he built a platform called Gradecraft and he said, Gamification is important, but it's not always intended to be fun, right? It's intended to work, but by developing your role and thinking about how you get ownership of that learning, you become invested in it. I think that's a lot of what you're seeing in the classes as they've...

Yu Chen: Yes, definitely, because they have a lot of freedom. That's the anatomy we want to introduce, including we encourage them to choose their own topic, and number two, come up with their own solution. And I think that another important part of their ownership part is we encourage them to bring their cultural assets and their community problems. So that's another motivation for students as well.

Vincent del Casino: Yeah, that's fantastic because I think that's where higher ed needs to be going, right, in order to really take our amazing, bright, thoughtful, engaging, and students who come from these very diverse backgrounds don't always have family who have an experience in college and give them an opportunity to own that. Is that then, so the scale of this now is with a number of universities through the research project, but this AI for social good project. Can you tell me a little bit more about what AI for the social good means and how the platform you're building is meant to try to scale some of this work?

Yu Chen: Yeah, so for me, AI for social good, meaning using AI as an application or tool to empower and address some of the most important social problems that can be locally or globally. And in particular, in this project right now, we are trying to align with the United Nations Sustainable Development Goals, where students, they are able to choose one of the goals among the UNSDGs. And see how that's manifested in their own community and address problems. And starting from 2022, when we fortunately get the funding from National Science Foundation, we started to go beyond SJSU. Right now it has been implemented in, beyond SJ SU, two other universities, CSU San Bernardino and Cal Poly Pomona, and actually two different domains. One is CS from... Of San Bernardino and the other is Geography from Pomona.

Vincent del Casino: I didn't actually know that. I'm a geographer by training. I actually hadn't, I knew it was a Pomona, but I hadn't realized it was in geography. I imagine around geographic information science sorts of questions.

Yu Chen: Yes, exactly. So when we're trying to propose that we didn't know how that might look like when that goes to different disciplines, I had little doubts about CS because AICS, they're kind of combined together and but for geography, we don't know. And the reason I invited our co-PI Dr. Gabriel Grankel from CPP is we met in another CSU events where He is working on using AI for agriculture. So thinking, oh, that's really cool. Maybe that might be related to your students as well. So last year, so right now, this is third semester we're implementing that. And actually in May this year, we showcased students AI for Sociology from our project in the first annual CSU AI for Social Innovation Symposium where we got to know how geography students, they are able to come up with their own AI prototypes using chat bot to address a diverse set of questions such as food insecurity, pollution, trash, etc. Within Southern California, which even though it's the same set of topic, very different from what our students come up with, even though the same. So that's really exciting to see.

Vincent del Casino: Yeah, I mean this is one of the moments where you see the power of the California State University as a system that you get to work with collaborators across. And we're talking like over a thousand students now who've been through some level of curriculum that must make you really excited about the potential.

Yu Chen: Yeah, definitely. I didn't look at the numbers until recently and say, well, actually, so many students have participated, including San Jose State, and not only MS students, but also other business faculties. And this semester, actually we have two more faculty from College business, one, Professor Matt McWa, and also Professor Mahesh Rajan, they are also implementing this into their management and marketing courses, making me to just, it's an again, an experiment to see how is that applying to students not known or traditionally thought about technology background. So yes, it is really rewarding to see the numbers to see, oh, those are the students who graduated from this project. And those are the numbers of projects that they came up with.

Vincent del Casino: So, what are the boundaries around the ethical questions, though? Because obviously, you're bringing that right into the curriculum. And I think what the opportunity is, if we train students with those kind of ethical principles, they're going to bring that out into the world. That's obviously a plan, but it is an interesting sort of question because technology firms and others who are kind of driving this tend to try to stay relatively neutral to the application question. But it does seem central to our kind of future as humans and to the planet and sustainability that we address these ethical questions. So how do you bring them into this curriculum tied to this question of what is the social good?

Yu Chen: Yeah, I think that I really love that question. And maybe I'll approach that question from two aspects. Number one is the standard way we're bringing AI ethics into the classroom. Basically, there are different approach that we're experimenting. Number one, it's a standard lecture style, teaching student what does it mean by data bias and how does that manifest in AI-related applications? And also, others include, for example, what does mean by the job market and employment and employ employment? And also what does it mean by privacy, security, et cetera. So those are the standard lecture we definitely inserted into our course components. And on top of that, there are different experimentation we're doing. Number one is students actually, if we have time during the semester before we give the lecture, students, they can do a debate in groups about the AI ethics before we gave them a lecture so that they get a chance to do a little bit of research. And then the third component is after students, they design their system, they will need to also propose and analyze what might be some of the risks of their design and then provide solutions, how to address them. So those are some of things that we designed to bring to the classroom. And some of other components that came from this project, which is a little bit pleasant surprise for us, is number one, when students, they are designing a system for diverse community, they get to know the diversity of the population in their community and whether there, there's any kind of inequity. Diversity issues in their community through there, they get to know more about why it is important to bring the fairness of data into their data set and system. That's the first pleasant surprise. And the other is, there are also some initiative about AI for social good in industry from the tech companies as well. For example, IBM, Microsoft, and Google, et cetera. And what we realize is that When we're thinking about AI for social good, it's important to think about who are we designing for and who are designing it and why. And the reason to ask those questions are important because, and again, tied back to the HCI mindset, if the AI for Social Good are designed by, for example, designers or developers who had less experience about what the users are actually going through or suffering. Then the system that designs may come from a good intention but might not be the good outcomes for people who are really suffering. So this is why we're trying to experiment this grassroots approach where a lot of innovation came from students who are coming from the community and I think that's another important aspect of AI ethics.

Vincent del Casino: So that's really, I could probably spend another two hours talking to you about this, because it's so interesting. But I want to just pull on one or two more things. One of which is the algorithms that undergird this and the ethical questions they collide in sort of, you know, how open can these systems be anymore? They're so complicated. You know, so in geography, in geographic information science, we, There are softwares out there that do open source, right? And open source is an interesting space. But the ability of the machines to kind of process is so intense, it's sometimes hard to get to that. And then the second question is how we're training these narrow and broad artificial intelligence algorithms to do the work we want them to do. And I think you bring that up, right, in the context of Should the developers be doing this or people on the ground who are communicating? How do we think about all of these issues knowing that the ability for machines to process is just surpasses our ability to do that now, like at the speed that.

Yu Chen: So I think right now, especially when people are all talking about chatGBT and generative AI, one thing we potentially need to think about, which we're also talking in the class, is the data sets from all over the internet, and how fair the data is, and where the data has generated from. And that will connect to another question, the digital divide, which already existed even before the age of AI. And when AI came, it's just going too fast. If there's no interventions going to widen the gap. So that's exactly why we're trying to engage more students, especially with diverse backgrounds, to join the conversation or even creative force of the whole movement. So in that way, they are not only contributing to data, but participating in the discussion, design, and creation of the system. And I think it's so important for the next generation of students right now. And the other thing is actually participating, they are also educating themselves and also their community members, their family members, from diverse backgrounds to know more about AI. So I think this is a ripple effect of how engaging students at large from a from diverse background to participate in this movement, which later I believe that can impact maybe some of the decision making or things. The trends with the industry community and university levels as well.

Vincent del Casino: So I find what you're doing just amazing and fascinating. I'm so glad to have the opportunity to sit down with you. So one kind of last question, is the generative AI space and the kind of anxieties that's produced at some level within education, not just higher education, but it's a world that we're already in. So the question that I have is kind of a follow-up to what you just said and as a way to kind of close out this discussion Where is teaching and learning going in the context of generative AI? And how do we clearly maintain the value of those human conversations and all of this and teach students to work collaboratively but thoughtfully and critically with generative tools?

Yu Chen: Yeah, definitely. So for me, there are, when generative AI and ChachiPT is coming so fast in the past year, number one, for this project, we want to get students exposed to the tools and also application and potential. On the other hand, we're intentionally making slower movement towards integrating because, and again, number one there's no clear policy and regulation available yet worldwide. So that's why we're cautiously moving forward. By integrating number one, the innovation component, number two, be more aware of the risks. So there are, I can talk about some of the experiments we're doing.

Vincent del Casino: Yes, I would love to know.

Yu Chen: Yeah, so first of all, I want to say I feel really grateful for SJSU, for this SJSU Adobe Digital Literacy Laboratory, which I was a part of the member, and there I learned about Adobe has their tool called Adobe Express that can allow students, users, to put their prompts and keywords and generate images, which is pretty much similar to Dali from OpenAI. Right. And then. I was thinking, I want your student to get exposed to. So one experiment we're doing is also inspired by Center of Faculty Development, is we put them in a group for warm up and team building. They each needs to provide keywords to describe their identity, their culture and community, and put them the keywords to describe themselves and also their identity. And then allow the generative AI to create images that represent them. And then as a group, they create a collage of their group members to get a sense of identity and belonging. And after that, so there are multiple purposes. Number one is they, of course, they use the generitive AI. Number two, they bring this kind of social and community aspect into that. And number three, they need students after that. They will... Writes a little bit reflection. How do you evaluate this tool? Is it working? Is it not working? Are there any concerns? So this is one of the mini experiments we're bringing to students in this class. And then talking about the broader sense, right now we're approaching very cautiously because a lot of the students are building their chatbots. And the issue with OpenAI or ChatGPT to address community issue is We want to be really sure that the answers that are generated from open AI or generative AI in general is valid, is fair, and is true, which there is no guarantee so far. And that's exactly why we're very careful at the moment to use them to address community issue because the social issues has a lot of, it's touching. A lot of the population that might be underserved, which might, and their data might not be represented well enough over the internet, in the public internet. So that's why we're approaching very cautiously. So tied back to the education. So I've been thinking about this for a while, but right now I'm just thinking, in addition to teach students the skills, for example, AI and generative AI. Which are important for their employment and career in the future, at the same time, the skills are evolving. It's also very, very important, even more important, to teach students, figure out innovative ways, to engage students, to touch the human aspect more, including their critical thinking, their innovative spirit, and also the empathy and leadership, etc. So With the advance of AI and generative AI, I feel it's more important to touch the human intelligence even more.

Vincent del Casino: That's awesome. Well, I want to thank you so much for the conversation and taking time out of your busy schedule to discuss all this with me. It's just been absolutely fascinating.

Yu Chen: Thank you so much for having me here. I really enjoyed this conversation.