AI Confidence Gap: What WIRED and Harvard’s Survey Actually Found

A survey discussed by WIRED found a striking gap in reported AI familiarity. More men than women said they recognized two fictional AI leaders. That response does not prove anyone deliberately lied. The research also points to differences in workplace training and who gets heard in AI coverage.

This content is blocked because it would connect to YouTube.
This content is blocked because it would connect to Spotify.

What the survey found

WIRED senior writer Kate Taylor reported on an October 14, 2026 Harvard Kennedy School Shorenstein Center study of more than 5,400 Americans. Researchers included two fictional names among people associated with AI. In WIRED’s account, 53% of male respondents and 40% of female respondents said they recognized those names. The gap is notable, but a survey answer cannot establish whether a respondent knowingly made a false claim, guessed, or confused a name with someone else.

The picture is more complicated than “one group knows AI and another does not.” Men were more likely to answer factual AI questions correctly, according to WIRED. Men and women answered incorrectly at roughly the same rate; women selected “not sure” more often. These are group-level survey findings, not descriptions of every individual.

Confidence is not the same as knowledge

AI discussions often reward the person willing to give a quick, certain answer. This study raises a more useful question: when does certainty reflect real understanding, and when does it simply sound persuasive? Someone who says “I’m not sure” may be showing caution rather than a lack of ability. Likewise, recognizing a name on a survey is not a meaningful test of practical AI literacy on its own.

For workplaces, a better measure is whether people can identify an AI tool’s limits, check its output, protect sensitive information, and decide when a human should take over. Those skills are learned. They should not be inferred from self-confidence or a single survey response.

Training and visibility shape participation

The Shorenstein Center reports that 46% of women versus 36% of men surveyed said they had never used AI at work. It also reports that 64% of women versus 49% of men had received no workplace AI training. Those differences suggest that access to tools and instruction should be part of any conversation about AI confidence.

The center also analyzed more than 70,000 AI news articles. In that sample, men were named in headlines 14.6 times as often as women and cited as experts 4.3 times as often after excluding executives, founders, officials, and investors. Media visibility does not by itself explain survey responses, but it affects whose expertise audiences regularly see. Broader training and a wider range of quoted experts could help more people participate in decisions about AI.

Watch and listen

The Short above gives the quick version; the podcast explores the confidence-versus-knowledge question and what the training gap means in more depth. Watch the YouTube Short and listen to the full Daily AI Chat episode. Sources: WIRED, Kate Taylor, October 14, 2026, and the Harvard Kennedy School Shorenstein Center’s study announcement.

Save to list

Scroll to Top