Men bluff more about AI knowledge than women

A recent survey by the Harvard Kennedy School, conducted among more than 5,400 Americans, reveals that more than half of men claim to know two fictional AI top executives. This percentage is much higher than that of women, of whom 40 percent claim the same. Participants were presented with a list of names, including well-known figures like Sam Altman and Mira Murati, but also with false names, such as Susan Milner and Craig Dalton. Remarkably, nearly 75 percent of respondents could identify both real names, while 53 percent of men claimed to know the made-up names as well.

The results not only suggest a difference in the level of confidence but also in behavior when knowledge is lacking. On average, men indicated 2.3 times that they didn't know, while women chose this option an average of 3.3 times. This indicates that men are quicker to take guesses, which is further underscored by their optimism about and confidence in AI.

Laura Manley from the Shorenstein Center explains that this phenomenon, where men are mentioned more often in the media, also plays a role. In a year filled with AI news, men appeared in headlines more than fourteen times as often as women, and three-quarters of the cited experts were male. This contributes to men's self-perception as AI experts.

A similar pattern has also been observed in the Netherlands. A study by IT company Conclusion showed that 76 percent of men considered themselves at least 'basic users' of AI, compared to 66 percent of women. Nevertheless, both men and women scored an average below par on questions about AI knowledge. Only 20 percent of participants were familiar with the term 'language model.' When asked about AI hallucinations, 51 percent of men could provide the correct answer, while only 32 percent of women identified this phenomenon.

Overall, it seems to be a social phenomenon that plays out not only in the AI sector but has also been observed in other disciplines, such as education among students. Men tend to be more confident in their supposed knowledge, even when that knowledge is often unfounded. This results in a search for more factual figures and insightful data as represented in AI statistics.

Read the full article from AI Insider.