Publicly available AI is reshaping how teams work

By Kevin Manne

BUFFALO, N.Y. — The public release of ChatGPT in 2022 marked a turning point for how people interact with artificial intelligence, and now scholars need to rethink what counts as a team, how those teams coordinate with AI and how to study those dynamics, according to a University at Buffalo School of Management researcher.

In an article forthcoming in a special issue of Small Group Research, co-author Kate Bezrukova, PhD, associate professor of organization and human resources in the UB School of Management, proposes setting a clear boundary for researchers between “pre-public AI” and “post-public AI” because many team members now use AI in everyday life and team theories need to account for this human-AI collaboration.

“Humans have had millions of years to evolve the ability to work in teams and develop group work skills, but now that AI is public and common, teams are no longer just people,” says Bezrukova. “As AI becomes an integral part of our technological environment, we must quickly and carefully reengineer our theories and methods to reevaluate group boundaries, dynamics and the very nature of teams.” 

The authors analyzed a range of prior studies to determine how AI influences teams and found that the type of AI a team uses can shape trust, attention and problem-solving. Other studies find AI can boost creativity for novices but may reduce quality when relied on too heavily by experts.

Building on a classic definition of teams (people with complementary skills, united around shared goals and accountable to each other), the authors say new considerations need to include who (or what) counts as a team member, what skills matter, how teams align and who is responsible. 

To keep up with AI’s rapid change of pace, Bezrukova says researchers must move quickly, and calls for collaboration between social and computer scientists, transparency about the AI tools used and their functionality, and methodological innovation to capture the evolution of human-AI teams.

“Never in modern social science have we seen such a significant shift at such a pace,” she says. “In the future, researchers will need to study these differences and the new stages we go through as our tools and knowledge continue to evolve.” 

Bezrukova collaborated on the paper with Terri Griffith, PhD, professor of innovation and entrepreneurship in the Simon Fraser University Beedie School of Business.