In a recent survey conducted among over 5,400 Americans, it was revealed that 73% of male respondents claimed to be familiar with Sam Altman, the CEO of OpenAI. However, only slightly more than half, precisely 53%, acknowledged knowing Mira Murati, Susan Milner, and Craig Dalton. The twist is that while Murati is indeed a real person—formerly the Chief Technology Officer at OpenAI and currently the CEO of Thinking Machine Labs—Milner and Dalton are entirely fictitious.
Laura Manley, the Executive Director of the Shorenstein Centre at Harvard Kennedy School, notes that “we observed men exhibiting more optimism and self-assurance regarding AI compared to women.” This confidence can sometimes border on overconfidence, as illustrated by the recognition of non-existent AI executives. To provide further context, 40% of women reported knowing who Milner and Dalton were, while only 38% recognised Murati.
Understanding the Confidence Gap in AI
Published on Wednesday, the report delves into the nuances of the confidence gap concerning AI and the implications of gender divisions. Researchers found that men tend to use AI more frequently and, on average, are more adept at answering objective questions about the technology. However, both genders display a similar rate of incorrect responses. The crux of the discrepancy lies in the frequency with which women expressed uncertainty in their answers: an average of 3.3 out of 12 questions compared to 2.3 for men.
This inclination towards certainty prompted researchers to examine whether it affects media coverage of AI. Over the course of a year, men appeared in AI headlines more than 14 times more often than women. Given that leadership roles in the AI sector are predominantly held by men, the researchers chose to re-evaluate the media coverage, specifically focusing on individuals cited as subject matter experts. Alarmingly, three-quarters of these experts were male, a figure that exceeds the approximately 29% representation of women in the overall AI workforce.
The Media’s Role in Shaping Perceptions
“Perhaps there should be more women prominently featured as experts or given platforms to voice their opinions; I’m not sure,” Manley comments. Despite the prevailing bravado, Americans generally expressed a sense of dread concerning AI, with both men and women identifying “concern” as their primary emotional response. Nonetheless, women reported feeling more anxious, fearful, and unsettled than their male counterparts, who, in turn, rated positive emotions—such as curiosity, excitement, and enthusiasm—higher than women did.
Current discourse, according to Manley, is rife with assertions of confidence, whether regarding apocalyptic scenarios or boundless potential, which tends to drown out voices that might express uncertainty. There remain numerous unanswered questions regarding the appropriate uses and future of AI. One gender, simply put, is less inclined to acknowledge that uncertainty.
The Need for Greater Acknowledgment of Uncertainty
“We need to be able to say ‘I’m not sure’ more often,” proposes Manley. This sentiment underscores the importance of fostering an environment where uncertainty can be openly discussed, allowing for a more balanced conversation about the implications of AI technology. Addressing the confidence gap not only benefits the discourse around AI but also paves the way for a more inclusive approach to technology’s future.