7 02 / 23 PERSONALquarterly I also think one of the big open questions is where you insert AI in a workflow. Right now most uses position AI doing work that then feeds into a decision being made by a human. But the reverse could also be the case, where people direct AI and then rely fully on its decisions. That would certainly mark a transition where AI is not just an assistant. PERSONALquarterly: How do teams working with AI differ compared to traditional teams, for example, in terms of goal achievement in teams? Leslie DeChurch: Teams working with AI differ from traditional teams in that there is an additional layer between people and intelligent machines. I would speculate that many of the same processes that are important in teams will continue to be important in human AI teams, but they might look a little bit different. For example, we know that developing transactive memory systems is one of the cognitive foundations that allows people to work together effectively. People need to understand who knows what, and coordinate their expertise. With human AI teams, transactive memory will also include a meta layer of people understanding the AI and understanding their teammates understanding of the AI. These mental perceptions are the foundation of effective teamwork, they enable human teams to work together. When we add intelligent machines to teams, cognitive processes will gain an order of magnitude of complexity. That said, I think much of what we know about teams in terms of the kinds of processes and states that are important to perform will be extremely helpful in understanding both how to design the technology and how to develop team members’ capability to use it. PERSONALquarterly: What do team members think about their AI „colleagues“? Leslie DeChurch: This is my favorite question of all! We‘ve been doing lab experiments where we use a Wizard of Oz technology to introduce an AI into teams and then we observe how three and four member teams who‘ve been working together for about an hour react to their new AI teammate. We‘ve done some studies manipulating the kinds of contributions the AI makes. What‘s been most interesting to me is that people respond extremely well and, in some cases, too well to an AI doing task work. I say too well because sometimes they get lazy and rely on the AI to just do the work, or they start disregarding their other human teammates contributions in favor of just having the AI do the task. PERSONALquarterly: Could we use AI to optimize processes in teams? What will happen when AI tries to regulate team social interactions? Leslie DeChurch Something unexpected happened when we had the AI help regulate team social interactions. There is an abundant literature on team process that finds that sometimes teams do not perform as well as they should because of reLESLIE DECHURCH Department Chair, Professor E-Mail: dechurch@northwestern.edu https://communication.northwestern.edu/faculty/lesliedechurch.html Leslie DeChurch is Professor and Department Chair of Communication Studies & Professor of Psychology at Northwestern University. She is an expert on leadership and team effectiveness. Her current research is situated in three contexts: human-machine partnerships, space exploration, and innovation. DeChurch holds a PhD in Industrial and Organizational Psychology, is the Past-President of INGRoup (Interdisciplinary Network for Group Research), and Fellow of the American Psychological Association (APA), Association for Psychological Science (APS), and the Society of Industrial & Organizational Psychology (SIOP).
RkJQdWJsaXNoZXIy Mjc4MQ==