According to recent research from Cornell University, a digital teammate powered by artificial intelligence and featuring a female voice significantly enhances participation and productivity among women in predominantly male-dominated groups.
Researchers have discovered that the gender of an artificial intelligence’s voice can significantly influence the interactions within gender-imbalanced teams, ultimately informing the development of bots designed to facilitate effective human-AI collaboration.
Findings align with previous research indicating that minority team members are more likely to participate when their organization includes colleagues who share similar demographics or experiences, according to Angel Hsing-Chi Hwang, a postdoctoral researcher in information science and lead author of the study.
Researchers Hwang and Andrea Stevenson, affiliate professor of communication, investigated how AI might aid gender-imbalanced groups by conducting an experiment involving approximately 180 individuals self-identified as male or female.
Each group comprised a single individual – either a woman or a man – paired with a fourth agent that took the form of a summarizing voice, offering both masculine and feminine tones that appeared on-screen to guide directions, prompt ideas, and manage timekeeping. While the bot’s capabilities seemed flawless, there was a crucial caveat – it required human intervention to function optimally. In a classic “Wizard of Oz” experiment in human-computer interaction, Hwang operated behind the curtain, feeding strains of text produced by ChatGPT to the AI-powered chatbot.
Following the experiment, Hwang and Gainas scrutinized the chat log transcripts of staff discussions to identify patterns in the way typical contributors presented ideas or supporting evidence. Furthermore, contributors were asked to leverage their collective knowledge and expertise in crafting their submissions.
“After analyzing the specific actions of our contributors, we noticed significant differences emerge between female and male participants’ reactions depending on whether a female or male team member was present.”
“One intriguing aspect of this study is that most participants didn’t express a preference for a masculine or feminine tone,” Gain’s quotation noted. “This implies that people’s social inferences about AI may remain influential, even when individuals don’t perceive them as crucial.”
According to a study, women in minority groups tended to engage more when interacting with AI voices having a feminine tone, while men in similar situations were more vocal but exhibited decreased focus on tasks when conversing with a male-voiced bot. According to researchers, women displayed significantly more positive attitudes towards their AI teammates when they were part of a minority group within the team, differing from the men’s perspectives.
The AI agent’s ability to adopt a gendered voice enables it to provide a limited extent of assistance specifically to women from minority groups when they gather together, according to Hwang.
