Hello! Welcome to this episode of the Architects of Communication Scholarship podcast series, a production of the ICA Podcast Network. I’m Ellen Wartella. Today, we’re hearing from Dr. Shyam Sundar. Dr. Shyam Sundar is a James P. Jimirro Professor of Media Effects and Co-Director of the Media Effects Research Laboratory at Penn State University. He is also the Director for the Center for Socially Responsible Artificial Intelligence.With his interdisciplinary knowledge of psychology, computer science, and engineering, Dr. Sundar offers a unique perspective on topics revolving around interactive media, artificial intelligence, fake news, and the future of human-computer relationships. He is an ICA Fellow, a prolific scholar, a passionate educator, and a keen mentor. Our interviewer today is Saraswathi Bellur, an associate professor in the Department of Communication at the University of Connecticut, where she studies the effects of interactive media and has worked alongside Dr. Sundar during her graduate studies at Penn State University. Here’s Saraswathi.
Thank you for doing this, I look forward to our conversation.
Hello and welcome to the Architects of Communication Scholarship podcast. I am Saraswathi Bellur. It’s my pleasure and honor to be talking with Dr. Sundar today.
Let's begin with your academic background, which has been eclectic. You did your undergraduate degree in engineering back in India, even as you pursued a side hustle as a journalist working for top newspapers. Your initial academic journey here in the US straddled both domains of psychology and communication, and eventually, you bring them all together. In a sense, your scholarship reflects this highly interdisciplinary training. Given this diverse background, what led you to become a communication scholar?
I started out as a student of engineering in my undergraduate years, and while my engineering school happened during the mornings, I was a journalist at night. I was in the newspaper offices taking assignments, going out and reporting and bringing back stories to print. In those days, my interests around media and journalism crystallized into an interest in audiences, and how they might be perceiving what we put out in our newspapers. Ultimately, I've always been really interested in how people process mediated communications. While the focus of much of my current work, is communication technology, my ultimate goal in this work has always been to discover how aspects of technology affect the way people perceive content conveyed via those technologies. For example, I studied the effects of interactive media or interactivity, a concept that you and I have worked on quite a bit. The dependent variables of some of this research is always about how people perceive that content, how that content shapes their worldview. A lot of my work has also been in the area of human-computer interaction. But that area, which is called HCI, is really a means to an end., and the end is really that I'm focused on is effectiveness of communication. Likewise, another area of interest for me, and in fact, my minor in my doctoral studies is psychology, which helps us understand the human mind. But in my work, it is again, a means to an end. To better understand how user psychology intersects with technology to affect communication. Ultimately, it's all about how people are affected by communication technology, and how it affects the nature of their communications, therefore; it's an obvious choice for me to be in this field, and no other field.
I must say Dolf Zillmann during my master's studies of the University of Alabama, really was the first mentor in this academic journey. I would say he turned me on to this whole area of media effects. I interpreted almost everything in terms of psychological effects. Even the tech research that I do about user experience or social responses are all the psychology research that I was exposed to by the likes of Amos Tversky, Al Bandura, David Rumelhart, Robert Zajonc, Lee Ross, and Mark Lepper, whose classes I had the privilege of taking during my doctoral studies at Stanford. In my doctoral work, my doctoral advisor at Stanford, Cliff Nass was highly, highly instrumental in shaping my research interests in psychology of technology, per se. My doctoral committee members, Byron Reeves, Steve Chaffee, and Don Roberts, were quite instrumental in helping me think about technological phenomena, in conceptual terms and proposed research frameworks that go beyond one study, and generate more study ideas.
The connection between the dots existed and now they're all coming together in the work that you do. When you entered this field, who were your mentors, and how did they shape your academic experiences?
and generate more study ideas,
that's
That's amazing. Standing on the shoulders of many giants. When you started your work in the area of communication technologies, theories, and models that examine technology-based processes did not exist. In fact, one of your major contributions to the field has been to put the medium square and center. Specifically, you have carved out a theoretical niche in the form of an affordance-based approach, building many models and theories, such as the Interactivity Effects Model, the MAIN Model, TIME Model, and extending it all the way to the field of AI more recently. So as a theorist, what are some issues and challenges that you have faced when it comes to theory building in a discipline that is continuously evolving?
This is a very interesting and important question, because as you know, technology changes all the time. If you build a theory around a given technology, then it will be outdated because the technology is going to be outdated. It's very important for technology theorists to be thinking about technology in conceptual terms, rather than in operational terms that are specific to one technology, and also in terms of building abstractions. Key challenge of doing, tech research and comm is that everyone focuses on content that is usually piped through a particular platform, and then they associate that platform with that particular content. They're focused on the content rather than appreciating the role, the overarching role of technological affordances. I've always tried to champion the cause of technological effects and user psychology, by building several disciplines, most especially communication, psychology, and computer science. In doing so, I've, over the years, delineated two broad types of technological effects. The first has to do with the effects of interface cues, which was published 13 years ago in the form of the MAIN model that you refer to. And the second, to the effects of user actions on the interface. What happens when users actually act on the interface instead of just being cued by the interface? Here, the focus has been on affordances, like interactivity and customization. The idea is that such actions will have psychological effects, and what people do or how the audience members, or users, as we call them, interact with the media can change the nature of the communication itself, and all the effects of that communication. These effects are no longer simply an effect of the features, but are a direct effect of the user actions on the interface and not simply the cues that exist on the interface. It's a culmination of different models that I've been working with and proposed in the preceding 20 years.
Most social scientists today are inundated with data thanks to big data and computational approaches toward understanding various social phenomena. Our ability to make sense of this data is far outpaced by the sheer volume of data available. How do you see the field rising up to this challenge where our existing theories may not be sufficient to explain what's going on?
This is a very current question and very important one. The developments in computational aspects have been much more profound than we even had the ability to anticipate. We are now at a stage where we are able to get more data about more specific things that people do with media than our theories envisioned. Our theories from the last 50-70 years, could even think about. So as a result, we have a big hammer in our hand, but we only see small nails. Sometimes we use the hammer to hit those small nails, expending a lot of effort to get at a minor point. We have this elaborate methodological toolkit that we deploy to make a very minor theoretical point, and so it ends up not being so parsimonious. But when you think about it in terms of the development of the field and evolution of scholarship, our methods have not just caught up to our theories, but really surpass them. We are now able to see more than we know what we are looking for. It's like we have a really powerful microscope. We can not only describe what we are seeing, and we still cannot make sense of all the things that are going on in that petri dish. There are so many unknown organisms that are unknown factors that are playing a role, but have not been accounted for by our current theories. That is why we need more homegrown theories that can catch up to the reality that we are able to capture with these advanced methods and help us make sense of what we are seeing with new powerful computational approaches. We now have data access to a lot more phenomena in the world. It is not surprising that computational data sciences are mostly descriptive at this stage. Why and how these phenomena occur can be explained by more homegrown theories. But first, we need to document all the things we are able to see with these new computationally intensive methods. While some scholars have sought to pursue description of media use, and content, generally others have adopted the strategy of studying each media user over time, and still others have focused on screen changes on mobile devices, second by second. All of us are in general trying to bring more structure to deal with this world of new data that is inundating us like you say. Simply thinking about all these data as a means to test theories that some psychologists came up some 30 years ago, I think doing a disservice to the richness of currently available data. Our theories are quite general, at this point. In a way, I would not blame the computational sciences for being atheoretical I would blame the existing theories for not being sufficiently nuanced for us to be able to entertain this level of our granularity in our data. So I foresee that in the next few years and decades, there will be a meeting of the two. The granularity of available data will inspire theorists to theorize about certain phenomena at that level and data will then fit the theory better. But right now, we have granular data and general theory and the fit is not so great. A simple example from my own work is the research tradition of uses and gratifications. Which has historically had, very broad labels like: information seeking, escapism, surveillance. That is pretty much true for all our online media activities. I find these broad gratifications deeply unsatisfying when applied to modern digital media because they do not explain why, for example, many adolescents use Snapchat, or why people switch from Facebook to Instagram, or why some people prefer textual platforms like Twitter, while others prefer more visual platforms like Pinterest. For me, those kinds of questions can be answered only if we take into account some of the newer gratifications that emerged from the newer aspects of these technologies relating to modality, interactivity, navigability, and so on. I feel the urgent need for more data-relevant theorizing in order to effectively leverage the power of the emerging computational sciences.
we are seeing with new powerful computational approaches.
I think you make an excellent point that I really like the microscope analogy,
I think you make an excellent point there. I really like the microscope analogy, because, yes, we have a lot of things that we're looking at, but if we don't have the generative frameworks and particular ways of thinking about that new data, it's not going to be as relevant or as meaningful. The questions that we ask have to rise up to the challenges that we face in terms of data.
We first have to describe what we see, so that the theorists can get an understanding of what is it that we are able to see with these tools before we can even begin to think about them theoretically. That's why you find a lot of the big data work tends to be on the descriptive side. This is part of the evolution of this type of research as it'll take a while before it becomes more theoretically meaningful.
The answers we find are only as good as the questions we ask, and it's time for asking the right questions.
Right, no amount of data, no amount of granular data is going to give us insights if the questions are not geared toward that level of data.
So you have deep roots in the field of psychology and media effects, and that's reflected in the experimental approach that you employ in most of your studies. However, your research has also been widely interdisciplinary, combining psychology communication, computer science, human-computer interaction, and more. Has this changed your methodological leanings and the type of research that you see emerging in the field right now?
I think we are all influenced by the different fields and audiences, to which, we cater to, in a way, when we submit our research to these different interdisciplinary venues. We have to invariably kind of alter the way we frame our work and the way we even study some of the phenomena of interest to these different communities. I would say that the more core scientific approach of hypothesis-driven research that we pursue in psychology and in communication, especially in my sub-field of media effects, will need to undergo a transformation. When I submit to, let's say, HCI, in the field of human-computer interaction, which is dominated by designers and engineers from the industry. For them, the hypothesis-driven research is not quite as appealing it doesn't resonate with them, as let's say, a problem-solving approach. I end up having to frame my research in terms of what is the problem here and how do we solve this problem? A lot of the communication questions are not based on some theory that is being tested, which is the theory hypothesis-driven approach, but rather, a problem has to be identified with the current human media equation, and then see if we can come up with some clever ways to solve that problem using research.
Looking at questions like why do we fall for fake news? Do we trust machines too much? How can we build better trust and credibility in our online interactions which have a strong practical implication in how they are examined? Looking forward, what do you think are some of the big societal challenges where communication scholarship can make a major contribution?
In general, I think the question itself sets up the response. When you talk about why do we fall for fake news? This has become a very big issue just in the last five or so years. It gets at the heart of communication in my mind, because communication, especially mediated communication, is nothing if it's not credible. Historically, media entities have risen and fallen based on the degree of credibility that they have with their audiences. Restoring information credibility by addressing the scourge of misinformation, is one of the biggest challenges and also an opportunity for our field for communication scholars, and to come up with different ways in which we can tackle this problem of fake news or false information, that is polluting the information environment. It's not just about fake news reducing our trust in these institutions, but it's also about us not being able to participate effectively in government, in politics, in our civic duties, and so forth, because we don't know how to differentiate between fake and real. There's a real need for media literacy, and I think that's a big challenge for people in our field, and also the contribution that our field can make, to even training students. Addressing this scourge of misinformation from many different angles from containing it to educating people about it, to coming up with algorithms to automatically flag them so that readers are more aware of them and discount potentially fake information. Another big societal challenge I see is with the arrival of artificial intelligence. You mentioned in your question about, do we trust machines too much? The answer is, yes. I've given talks about this for several years now, and the reason this topic has even come up is because machines have commanded a whole lot of respect from users, for a long time now, because machines are considered infallible, they command a lot of respect. They have high levels of credibility to begin with. AI machines are taking on much more agency, and as a result, they're undermining the agency of the user. While we appreciate the convenience that machines offer, we appreciate the personalization that they offer, we are also very often scared by the surveillance aspects by how much they invade our privacy, and how much control they might take away from us. I think the future lies in good ways to build frameworks for effective collaborative work between humans and AI, how we can move forward where both machines and humans co-construct reality by sharing agency, rather than asserting or giving up agency.
We have seen where you have mentored hundreds of students, aspiring scholars, and media practitioners, and having been a direct beneficiary of Shyam’s lab group at Penn State, I can attest to how a collaborative research and mentoring approach can work wonders. From your experience tons of running this lab group so successfully for many years now, what do you think are some of the main qualities that contribute to a successful mentorship both for the mentor and the mentee?
Lab group is one of those things that I acquired from another field, from the engineering and human- computer interaction domains, where I'm active in. Right around the time that you joined Penn State and became my advisee I decided to make the switch from an apprentice model to a lab group approach where all my advisees work together, and they meet every week, once a week, all 12 months, actually, with some breaks, here and there, as you know. The reason we do this is to organize ourselves. I like to joke sometimes that lab group is to research what mafia is to crime, it's organized. It's a good way to have public accountability and organize all our research there. Everybody is responsible for some aspect of the lab group's work. It's also a committee of peers. That's what the lab group does is bring together people who are all in the same boat. Grad school is a lonely enterprise as it is, and I feel like having a lab group provides a much-needed sense of community, and peer consultation that goes on and the mutual learning that comes from it. Lab group mentality itself is something that I tried to inculcate in my students.
I think it does take a village to raise and groom a scholar. As you mentioned, just the joy of learning together and exploring things together and witnessing the growth that happens both at an individual level and collective level is something amazing. So since this podcast series is titled, Architects of Communication Scholarship, what would you say you've built?
I don't know if I built anything specifically, but I'd like to think of myself as a champion for technology effects. I've been championing the cause of technological effects and user psychology, and also championing interdisciplinary work, or multidisciplinary work, bringing several disciplines most especially communication, psychology, and computer science into the fold and trying to understand them together. My main work has been in the area of the effects of technology. The first type of effects that I spent a good deal of early part of my career was on the effects of interface cues. Which was published some 15 years ago in the form of MAIN model. The second area pertains to the effects of user actions on the interface with a focus on interactivity and customization. The idea is that such actions have psychological effects, what people do, or how the audience members, or users, as we call them, interact with the media can change the nature of communication itself, and all the effects of that communication. These effects are a direct result of user actions on the interface, and not simply those of cues on the interface. I pulled them all together, under the Theory of Interactive Media Effects, or TIME. It's a culmination of different models that I've been working on. Some of those models are: Interactivity Effects Model, the Agency Model of Customization, the Motivational Technology Model, of which you're a co-author. The Motivation Technology Model is really premised on self-determination theory to explain how people can be motivated to engage with technologies, especially, that have pro-social outcomes, like health apps. Whereas the Agency Model of Customization is the psychological effect that happens when you customize your media environment. Or when you create or curate content, which is increasingly possible with modern social media. The Interactivity Effects Model has to do with delineating the different kinds of interactivity that we have, and typologizing. This is a typology that I came up with. Even before this, during my PhD dissertation days, I came up with a typology of online sources. For me, the confusing multiplicity of sources that came about with the arrival of the World Wide Web, where we could not tell if the source was a newspaper, or it was another person, or if it was other people on the internet, or serving as sources of news. For me, that presented a big challenge for the reader because the reader cannot factor sources into their evaluation of stories. The source signal was becoming much more muddied than it was with television or newspapers. I needed to figure out a strategy to typologize, and that's really what I did for my dissertation. Then more recently, I've come up with different typologies of uses and gratifications so that they can be applied to modern media. In many ways, I see myself as initially a conceptualizer, or I've been conceptualizing things and putting them in different buckets. That's what the work of typology is, and building taxonomies, and then graduated to building models, and then graduated them to building a theory, the Theory of Interactive Media Effects. Then, that theory has been, thankfully quite generative, in that I've been able to apply it to, for example, AI. Resulting in the Human AI Interaction Model, which is what we call the HAII TIME, that's the study of human-AI interaction from the perspective of the Theory of Interactive Media Effects. These I would say, broadly, are my contributions to the field.
It's amazing. You've added several tools to the theoretical toolkit that has empowered the whole scope of research that a lot of communication technology researchers engage in. As we wrap up this wonderful conversation here, what advice would you give to students and young scholars about their careers and life in academia in general?
The one thing that I tell all my students, and the one thing I look for when I vet candidates for positions in my lab, or even positions at our university, I think a key hallmark of a good academic is intellectual curiosity. That they are driven by the curiosity to dig in, and have the intellectual stamina to keep digging in and pursuing a particular question with great depth and rigor. You’ll see that when a scholar presents their work, they are very programmatic in the way they have approached the different studies. They're all threading a needle if you will. Pointing toward a grand question of great intellectual merit and societal benefit, rather than being opportunistic, doing one study here, one study there that really is not cobbled together well. Another key characteristic for me is passion. I think this work is not conventionally rewarded. It's if you get into this business, you need to be driven by some intrinsic motivation, and just the joy of doing this. You have to be passionate about your area of research. I think that is something that we are seeing an upsurge in especially now. We have so many public scholars and activist scholars, who are driven by particular ideas about changing the way the world works, changing the role of communication. I think we are at the cusp of a big paradigm shift in the way we think about the role of communication in society. I think that's, for me, a signal of passion, more than anything else. I think, finally, it's very important for young scholars to have fun doing it. I wouldn't be doing this now, I'm now in my 27th year as a professor, and I pretty much do what I did when I first came is conduct studies, work with my grad students, publish them. I wouldn't be doing this if it were not fun. Finding that fun, I think, is an important aspect of being a successful scholar.
Wonderful Shyam. Intellectual curiosity, intellectual stamina, finding your passion, and having fun; I think those are some excellent takeaways. As always, it's been a pleasure, Shyam, chatting with you. I learned so much, I'm inspired, and on behalf of everyone listening to this ICA podcast, and the communication community, thank you so much for your time and your insights.
Thank you, Saras for being such a great mentee and for doing a wonderful job of asking me all the right questions and for all your insights over the years as well. I really enjoyed working with you and I certainly enjoyed this chat that we had today.
Thank you, Shyam.
Architects of Communication Scholarship is a production of the International Communication Association Podcast Network. This series is sponsored by The School of Communication at Hong Kong Baptist University. Our producer is Bennett Pack.. Our executive producer is DeVante [Dee-Von-Tay] Brown. The theme music is by Humans Win. For more information about our participants on this episode, as well as our sponsor, be sure to check the episode description. Thanks for listening.”
