Cornell researchers report that women team members in a recent study tended to participate more when a AI partner also sounded female.
With the modern corporate world’s focus on ultra-productivity, many researchers are looking into ways to optimize team dynamics. In one study, scientists found that teams with equal numbers of men and women (the only genders tested) performed better in sales and profits than their male-dominated counterparts.
However, equal gender in a team is not always the norm, as men make up a majority of leadership positions in larger companies. The number of women and other underrepresented groups in the executive suite of companies has risen from a pitiful 17% to 28% since 2015.
Now, scientists from Cornell University may have a possible solution in the form of a “gendered” AI team member. In a new paper, Cornell researchers found that women team members tended to participate more in team activities and decisions if one of their team members, the AI partner, sounded female rather than male.
“Adding agents [AI team members] to the teams can make similar influences on team dynamics as adding new human teammates to human-only teams,” Dr. Angel Hsing-Chi Hwang, a postdoctoral associate at Cornell University and the study’s lead author, told The Debrief. The findings Hwang and her colleagues found mimic previous psychological research, showing that minority team members will participate more if there are team members who are like them.
Partnering with an AI Teammate
In the study, Hwang and her collaborators divided 180 men and women into groups of three and asked each group to virtually work together on a project (the study only included individuals who identified as male or female).
Each team had at least one man and woman and a fourth AI partner, programmed to sound male or female. This AI team member handled timekeeping, read instructions, and contributed the occasional idea. However, this team member wasn’t entirely automated, as Hwang fed dialogue lines generated by ChatGPT into the AI team member’s programming.
After the experiment ended, the researchers analyzed the chat dialogue boxes from each team to examine participation and productivity levels. They also surveyed each participant to understand their experiences.
“One interesting thing about this study is that most participants didn’t express a preference for a male- or female-sounding voice,” elaborated Andrea Stevenson Won, an associate professor of communication in the College of Agriculture and Life Sciences and a co-author of the paper, in a recent press release. “This implies that people’s social inferences about AI can be influential even when people don’t believe they are important.”
The researchers found that when a male individual was in the minority (working with two female team members), they tended to be more talkative but less preoccupied with individual tasks. In contrast, if a female team member was in the minority, they participated more when the AI team member sounded female.
According to Hwang and Won, the female participants felt they had a much more positive experience in the minority with a female AI teammate.
“In our study, we saw that having female voice agents in gendered-imbalanced teams could encourage minority female teammates to contribute more ideas during team discussions,” Hwang told The Debrief. “In that regard, we did not observe the negative impact of having gendered agents as teammates in our study. However, the positive effect of having gendered agents on teams remains rather limited on participants who have experienced severe marginalization in real life.”
The Importance of Representation
Whether it’s an AI team member who sounds female or a female colleague, the results of this study echo a broader trend: the representation of women and other underrepresented groups will encourage more participation from these groups. This is often due to these individuals being able to see someone like them succeeding in a specific field, making it seem more achievable and real.
This can be especially key in fields more historically dominated by men, such as engineering, science, and technology. While hiring more individuals from minority groups can help boost productivity (as mentioned in the study above), retaining these individuals can be difficult without a healthy and supportive culture. Using tools such as female-coded AI assistants to make these individuals feel more welcome can be beneficial in ensuring maximum group productivity and creativity for companies.
For Hwang, the next steps are to go beyond gender and look at other qualities.
“As mentioned, gender is only one of the many, many ways to represent one’s identity,” she said. “I would be curious to see how agents represented with other identity labels might influence team dynamics.”
Kenna Hughes-Castleberry is the Science Communicator at JILA (a world-leading physics research institute) and a science writer at The Debrief. Follow and connect with her on X or contact her via email at kenna@thedebrief.org