Paragraph 1
Let’s be honest: if you’re the kind of person who reads WIRED, you probably think you’ve got a decent grip on the world of artificial intelligence. You can rattle off the names of the usual suspects—OpenAI’s Sam Altman, maybe Demis Hassabis, perhaps even a few researchers you follow on X. But here’s a slightly uncomfortable question: when was the last time you actually paused and admitted you didn’t know something? Not just in AI, but in any field where you consider yourself well-informed? For many of us, especially men, that admission is rare. It’s not that we’re lying; it’s that our brains are wired to fill in gaps with plausible guesses, and sometimes those guesses sound a lot like facts. That’s the human condition. And according to a new survey from researchers at Harvard Kennedy School’s Shorenstein Center, this tendency is playing out in a fascinating and slightly alarming way when it comes to public perceptions of artificial intelligence. The survey, which polled more than 5,400 Americans, asked people about their familiarity with AI leaders. The results might make you chuckle—or cringe. When asked if they recognized Sam Altman, a huge 73 percent of men said yes. That’s not surprising; he’s practically a celebrity in the tech world. But then the researchers threw in two names that don’t exist—Susan Milner and Craig Dalton. And here’s the kicker: 53 percent of male respondents claimed they recognized them, alongside a real person, Mira Murati, who genuinely exists as the former CTO of OpenAI and current CEO of Thinking Machine Labs. It’s a perfect, almost painfully ironic demonstration of confidence overriding accuracy. But before we gleefully point fingers, it’s worth understanding that this isn’t just about vanity or weird survey tricks. It’s a window into how gender shapes our relationship with technology, uncertainty, and even the future itself.
Paragraph 2
The numbers from the survey paint a clear—and slightly uncomfortable—picture. Women, it turns out, were far less likely to claim familiarity with the fictional names. Only 40 percent said they recognized Milner and Dalton, and 38 percent recognized Murati. But the gap isn’t about knowledge; it’s about a very human tendency to confuse confidence with competence. When the researchers dug deeper, they found that men are, on average, better at answering factual questions about AI. They also use AI tools more frequently. But here’s the twist: both men and women get the same number of questions wrong. The real difference lies in how often they choose to say “I don’t know.” Women, on average, admitted uncertainty on 3.3 questions out of 12, while men did so on only 2.3. That half-point gap might seem small, but it represents a fundamental difference in how we handle ambiguity. In a field as new and rapidly evolving as AI, where even experts debate the basic risks and benefits, saying “I’m not sure” is not a sign of weakness—it’s a sign of intellectual honesty. But in a culture that often rewards certainty, especially in men, admitting doubt can feel like a failure. Laura Manley, executive director of the Shorenstein Center, puts it succinctly: “We’re seeing that men are more optimistic and more confident across the board when it comes to AI.” That confidence can be productive; it drives innovation and risk-taking. But it can also veer into overconfidence, precisely the kind that leads someone to nod knowingly at a name they’ve never heard. And overconfidence has consequences. If we make decisions based on what we think we know rather than what we actually know, we might embrace AI in ways that are premature, or dismiss concerns that deserve serious consideration.
Paragraph 3
So what does this confidence gap actually look like in the wild? Let’s consider how AI is being integrated into daily life. Imagine a manager deciding whether to trust an AI tool for hiring. If that manager is a man who reads a lot of tech headlines, he might feel he has a solid grasp of the system’s capabilities. He might not realize that his confidence is built on a shaky foundation of half-remembered articles and optimistic guesses. His female colleague, meanwhile, might be more cautious, asking pointed questions about bias and data quality. She might not have the same level of self-assurance, but she might be more accurate about what she doesn’t know. The survey data supports this: while men answer factual questions about AI correctly more often, they also make the same number of errors as women. The key is that women are more likely to acknowledge when they’re unsure. That’s not a flaw; it’s a feature. In a domain where new models and breakthroughs appear weekly, the ability to say “I don’t know” is essential for learning and for building robust, safe systems. Researchers found this pattern persists even after controlling for education and technical background. It’s not that women are less knowledgeable; it’s that they are less willing to project false certainty. And this has ripple effects far beyond the individual. When we make public policy, business decisions, or even family discussions about AI, we rely on a chorus of voices. If that chorus is dominated by people who are confident but not necessarily more accurate, we risk making choices based on bluster rather than evidence.
Paragraph 4
The media, as always, plays a huge role in amplifying this dynamic. The survey’s authors looked at a year’s worth of news coverage about AI and found something stark: men were named in headlines about AI more than 14 times as often as women. At first, that might seem like a simple reflection of the industry—top AI executives are overwhelmingly male. But when the researchers narrowed their focus to people quoted as subject experts—not just CEOs or politicians, but those brought in to explain the technology—they found that three-quarters of these experts were men. That’s a proportion far higher than the actual AI talent pool, which is roughly 29 percent women. So the imbalance isn’t just a story about who leads companies; it’s a story about who gets to define the narrative. When we see only men speaking with authority about AI, we absorb a subtle message about who belongs in this field and whose opinions matter. And when those men speak with unshakable certainty—whether they’re predicting utopia or apocalypse—they drown out the more nuanced, hesitant voices that might be asking tough questions. Manley notes, “Maybe there should be more women prominently featured as experts or given the platform to express: not sure.” That’s a crucial point. It’s not that women need to be more confident; it’s that our public discourse needs more room for uncertainty. A journalist might feel pressure to get a quote from a prominent male figure because he’s accessible and opinionated, but that doesn’t make him a better source. In fact, it might make him a worse one.
Paragraph 5
Beyond the questions of confidence and media representation, the survey also revealed telling differences in emotional responses to AI. Perhaps unsurprisingly, Americans overall said their top emotional response to AI was “concern.” But when broken down by gender, women were significantly more likely to report feeling concerned, fearful, and worried. Men, on the other hand, rated every positive emotion—curiosity, enthusiasm, excitement—higher than women did. This isn’t necessarily a bad thing; a little fear is healthy when a technology is this disruptive. But the split reveals a deeper cultural problem. In the current discourse, says Manley, “there’s so much confidence in either doom or abundance.” The loudest voices on both sides—the ones who say AI will save us and the ones who say it will destroy us—share a refusal to entertain uncertainty. That leaves very little oxygen for the middle ground, where most of us actually live: cautiously excited, moderately worried, and genuinely uncertain about what comes next. Why is it so hard to find people willing to say “not sure”? For men, social conditioning might play a role; they’re often taught to project strength and decisiveness, even when they’re not sure. For women, the opposite might be true: they’re socialized to be cautious, to double-check, to hedge. Neither extreme is ideal. But in a field as genuinely unpredictable as AI, humility is not just a virtue; it’s a necessity. We are making decisions today that will shape the next century—how we work, how we learn, how we govern. To make those decisions well, we need to acknowledge how much we don’t know.
Paragraph 6
So what’s the takeaway? It’s tempting to read this survey and laugh at the men who confidently recognize fictional AI executives. But the problem is far more systemic than a few embarrassing survey responses. We have built a culture—in tech, in media, in society—that rewards certainty and punishes doubt. We celebrate the visionary who promises to change the world, not the engineer who raises her hand and asks, “But what if this fails?” The result is that we’re steering a massive, powerful ship while many of its most thoughtful passengers are too shy to speak up. Manley’s plea is simple and direct: “We have to be able to say ‘not sure’ more often.” That means creating spaces where admitting uncertainty isn’t seen as weakness, but as wisdom. It means inviting more women and other underrepresented voices into the conversation, not just as tokens, but as experts who bring a different perspective on risk and possibility. It means changing the way we ask questions—in surveys, in news interviews, in boardrooms—so that “I don’t know” is a legitimate answer, not a disappointing one. And it means, for those of us who consider ourselves knowledgeable, doing a little more self-reflection before we nod along to a familiar name. The next time you hear about a new AI breakthrough, or an alarming prediction, or a startup founder you’ve never met, pause. Ask yourself: do I actually understand this, or am I just confident? And then, maybe, be brave enough to say, “I’m not sure.” Because in the grand, uncertain journey of humanity and artificial intelligence, that might be the most honest—and most useful—thing any of us can say.
