How working with AI impacts teams’ collective attention

The effectiveness of teamwork in professional and academic environments generally depends on several factors, including communication and coordinated attention. In this context, collective attention implies the ability of team members to focus cooperatively on the task at hand, optimally dividing it into subtasks and alternating their efforts to complete it as efficiently as possible.

Researchers at Northeastern University recently conducted a study to determine whether introducing artificial intelligence (AI)-based virtual assistants to a team of humans influences the team’s collective attention. Their findings, prepublished on arXivsuggest that using AI-based tools in a collaborative group can significantly influence what is discussed by team members and how it is discussed.

“AI is going to play an increasingly important role in how we work in the future,” study co-author Christoph Riedl told Tech Xplore. “Researchers have been studying dyads of a human working with an AI system for some time. Yet less research has looked at how entire teams will work with an AI system.”

The main goal of Riedl and colleagues’ recent study was to better understand how the presence of AI impacts teamwork. Rather than focusing solely on the dynamics between AI and individuals, they studied AI’s influence on the dynamics between all team members, including the dynamics between individuals.

“We conducted a lab experiment in which groups of 3-4 people were asked to solve a difficult puzzle,” Riedl explains. “To help them, they were given access to an AI assistant. During the experiment, the AI ​​assistant gave the team several hints on how to solve the puzzle. We recorded the team’s communication and looked for changes in the team’s communication pattern just before the AI ​​assistant gave the hint and just after.”

To determine whether a team’s communication was affected by the presence of an AI agent, the researchers carefully transcribed all recorded conversations between team members. They also noted the general topic of discussion at different times and the specific words used.

“We found two interesting things,” Riedl says. “First, the teams were heavily influenced by the AI ​​assistant in their speech. As soon as the AI ​​assistant presented a new idea, the humans immediately picked it up and talked about it, even if it was a poor-quality clue that didn’t help them solve the puzzle. Second, we observed that the humans in the teams adopted the language used by the AI ​​assistant.”

Experiments conducted by Riedl and his colleagues suggest that, when used during group work, AI assistants can influence both what is discussed by human team members and the specific terminology used. For example, the researchers found that while solving the puzzle used in their experiments, members of some teams initially referred to a specific symbol as “diamond,” but they began calling it “gem” after the AI ​​assistant did so.

Interestingly, this process of language adaptation appears to be automatic and occurs regardless of what participants think about the AI ​​assistant. For example, some teams began using the same language as the AI ​​even though they said they did not trust it and doubted its skills.

“I think the most notable implications of these findings are that when you introduce AI systems, there can be all sorts of unintended consequences,” Riedl said.

“In our case, the AI ​​system was designed to improve team performance through the cues it provides, but beyond that goal, the system subtly altered the teams’ collective attention and the language the teams used. Some of these impacts may have positive effects, such as better cognitive alignment between human team members, but others may have negative effects, such as diverting attention from the most relevant aspects of the task.”

The team’s recent work highlights the complex ripple effects that can occur when AI systems are introduced into professional and collaborative environments. While anticipating all the ways AI will affect the people it interacts with is a very difficult, if not impossible, goal, Riedl and his colleagues plan to continue studying the effects of these systems in a variety of settings.

“Our research shows that adopting a collective intelligence perspective and building on existing research in this area can give AI researchers an important foundational understanding of group dynamics,” Riedl added.

“Overall, this suggests that designing useful AI systems may be much harder than we thought, because designers now need to pay attention to how their system might affect other aspects of team dynamics. Currently, my lab is working on related research on the unintended consequences of AI. How does AI affect outcomes beyond “performance,” such as attention, learning, and organizational culture?”

© 2024 Science X Network