Should AI Be the Newest Member of Your Team?

The often-quoted adage, “Teamwork makes the dream work,” may be true, but there is still one important question missing from the equation. What makes the team work better?

That question is at the heart of Stevens School of Business Assistant Professor Bei Yan’s research.

“My research is really interested in how people use new technologies to collaborate and work together,” she said. “I come from a communication background, and I started with how small teams in an organization communicate with the support of technologies and how they work together to achieve a task.”

Dr. Yan’s work includes technology-supported collaborations and how groups are influenced. She has studied how small teams interact with intelligent personal assistants (IPAs) and the role of bots in large online crowdsourcing efforts such as Wikipedia. She experiments with and analyzes behavioral trace data of online communities. She describes her work as “the intersection of communication, social psychology and human-computer interaction.”

“I got interested because one of the things that is changing how people work together really is online collaboration and peer production,” Dr. Yan said. “One area of my research studies how peer production works because this is a new organizational form. There's less of a top-down managerial bureaucracy, but people manage to coordinate and produce amazing things like open-source software or Wikipedia together with the support of information technologies. These communities are also collectives in which bots are extensively deployed.”

Dr. Yan’s research has earned several awards, including the Best Paper Award at the Meeting of the Academy of Management in 2020 and 2023, the Dennis Gouran Research Award in the Group Communication Division from the National Communication Association (2020), and the Top Paper Award at the International Conference on Global Cultural and Creative Industries in 2019.

Her paper, “Stigmergy in Open Collaboration: An Empirical Investigation Based on Wikipedia,” was especially meaningful because it was her first collaboration with Stevens faculty and students. The article co-authored the article with Nico Zheng, a Stevens Ph.D. program graduate and current faculty member at UMass Boston, and Stevens colleagues Feng Mai and Jeff Nickerson, examined how participants in open collaboration communities are able to coordinate and co-produce knowledge without explicit communication, a process known as "stigmergy." By analyzing Wikipedia data, they found that higher degrees of spatial-temporal clustering of contributions, a new measure of stigmergy, were associated with greater community participation and higher quality knowledge production, helping practitioners design more sustainable open collaboration communities.

“I study what happens during the collaboration process,” she continued. “How are people using the technology? How do they interact with each other? How are these processes related to the outcomes? Do they form a shared mental structure so they can work together better in the future? How well do they perform the task?”

Learning to Work with Machine Learning

As artificial intelligence has become more accessible and prevalent, Dr. Yan’s focus has shifted to learning how people can integrate machines into their work to produce better results.

For example, the paper, “It Depends on the Timing: The Ripple Effect of AI on Team Decision-Making,” sponsored by the National Science Foundation, explored how the timing of AI assistance impacts team performance in decision-making tasks. Through online experiments involving new product development, Dr. Yan and her co-author found that teams receiving AI help early in their process made better decisions than those getting help later. Early AI assistance appeared to promote more information sharing among team members, suggesting a positive ripple effect on team dynamics.

“There are design implications for online platforms set up to support people working together,” Dr. Yan said. “For example, Microsoft and Zoom are creating and enabling intelligent agents to join meetings, work with teams and offer support. What is often understudied is how the teams would react. It's typically not known how the technologies interfere with their collaboration. The majority of the studies looking at AI study how an individual would interact with ChatGPT, for example. In teams, research in the collaboration process is more complicated. When people work together you would think they can utilize their diverse knowledge and create better things together, but what is often observed is they interfere with each other and sometimes come up with worse ideas than people working individually. Now the question is when the commercial applications of technical agents are pushed into the team context are they really going to help the team?”

Know Thyself

Ironically, Dr. Yan’s focus on teamwork and group settings started because she is, “not very good working with others.” That self-realization led her to think about the reasons why she often felt collaboration could not meet the desired outcomes.

“I think I'm good working by myself, so I started to be interested in teams because I feel when I work as part of the collective I start to change,” she said. “It’s hard to work with others, and I often think it doesn't really achieve the outcome that it is capable of. I started to think about why that happens. I started to be interested in technology because I realized that it has made lots of things that seemed to be impossible, possible. For example, having thousands of volunteers from around the world collaborate together on the Internet and create a very complicated encyclopedia like Wikipedia that everybody uses. This is a question that amazes me.”

Studying how teams work together might seem to fit more into the realm of communication studies or psychology, but Dr. Yan has found a home in the School of Business.

“My work could go in different places, but I like being in information systems in business the most,” she said. “What I do is very interdisciplinary. I was trained more as a social psychologist, but I study technology, online platforms and big data sets. A lot of my research overlaps with computer science, human-computer interaction and engineering design. I can use social psychological perspectives to study phenomenon that is more kind of human-computer interaction, but I'm producing theory that is more information systems based. I came to Stevens because I cited my colleagues’ work, and I felt like we share similar interests. But now I really enjoy being in information systems.”

The Road to Hoboken

Dr. Yan earned a bachelor’s degree in marketing management from Renmin University of China in 2010. She enrolled in a dual-degree program, completing her master’s in global media and communications from the London School of Economics in 2011 and in global communication at the University of Southern California in 2012. After deciding to pursue a career in academics, she stayed at USC and completed a Ph.D. in communication.

“When I was an undergraduate student, I wasn't clear whether I wanted to be an academic or not,” she said. “Gradually, I switched to research because I questioned if what I learned is enough. I want to know more things. I want to create more knowledge because there are things I don't understand, and I don't think there is existing knowledge about them.”

After finishing her doctoral work, Dr. Yan was a project scientist at the University of California, Santa Barbara until 2020 when she joined the Stevens faculty.

“I was very impressed with the focus on technology. I have a very strong research interest in technology and how technology works to shape human interaction,” she said. “That really attracted me to Stevens. Also, when I talk to my colleagues, I realize they're very open to interdisciplinary topics and collaboration, and I feel Stevens supports that. They also have diverse backgrounds, and they're very friendly.”