Do Men Really Know More About AI—or Just Sound More Certain? WIRED Examines a Harvard Survey on Fictional Experts, the AI Confidence Gap, and Who Gets Heard in Tech

SPEAKER_1 (0:09)

Welcome to The Daily AI Chat. This podcast was curated by our human friend Fred, and we are your dedicated AI hosts.

SPEAKER_2 (0:16)

Glad to be here today.

SPEAKER_1 (0:17)

So, put yourself in a scenario that you have probably experienced at some point in your career. You are sitting in a strategy meeting.

SPEAKER_2 (0:24)

Oh yeah, we've all been there.

SPEAKER_1 (0:25)

Right. Stakes are reasonably high. The coffee is definitely wearing off.

And someone across the table casually drops a name into the conversation. And they deliver it with this absolute unshakable authority. Instantly, you get that hollow feeling in your stomach because you have absolutely no idea who this person is.

SPEAKER_2 (0:44)

It's the worst feeling.

SPEAKER_1 (0:45)

It really is. You glance around the room. Half the table is nodding thoughtfully. The person next to you is like jotting something down.

SPEAKER_2 (0:50)

Just aggressively taking notes.

SPEAKER_1 (0:52)

Exactly. So, what is your move?

Do you stop the meeting, raise your hand, and ask for clarification, which effectively broadcasts that you might be completely out of the loop?

SPEAKER_2 (1:02)

Which nobody wants to do.

SPEAKER_1 (1:03)

Nobody. Or do you deploy the classic corporate survival reflex?

You furrow your brow, you nod along, and you project the illusion that you are completely tracking with a conversation.

SPEAKER_2 (1:15)

The pressure to just blend in with the consensus in those moments is enormous.

SPEAKER_1 (1:20)

Absolutely.

SPEAKER_2 (1:20)

We are just socially conditioned to protect our perceived competence, particularly when the subject matter feels really urgent, nobody wants to be the ones slowing down the momentum to ask a foundational question.

SPEAKER_1 (1:33)

There's arguably no subject matter projecting more urgency right now than artificial intelligence.

SPEAKER_2 (1:39)

Spot on.

SPEAKER_1 (1:40)

The pressure to appear fluent in AI is completely reshaping how we communicate at work, which actually brings us to the data driving today's deep dive.

SPEAKER_2 (1:48)

I am so excited for this one.

SPEAKER_1 (1:50)

Me too. We are looking at a really comprehensive new report by Kate Taylor for Wired. It was published today, actually, October 14, 2026

SPEAKER_2 (1:58)

Fresh off the press.

SPEAKER_1 (1:59)

Literally today. She is covering a massive freshly released study out of the Harvard Kennedy School Shorenstein Center.

SPEAKER_2 (2:05)

A really fascinating piece of research.

SPEAKER_1 (2:07)

It is. Our mission today is to unpack a really profound gender gap in how people perceive, talk about, and claim expertise in AI. We are asking the big question, in the modern enterprise, is raw confidence being actively mistaken for actual AI literacy?

SPEAKER_2 (2:25)

Right. But before we unpack the numbers in Kate Taylor's report, we do need to establish some analytical boundaries for you listening at home.

SPEAKER_1 (2:32)

Good call. Give us the guardrails.

SPEAKER_2 (2:34)

The researchers at the Shorenstein Center surveyed an extensive pool of more than 5,400 Americans.

SPEAKER_1 (2:39)

That's a huge sample.

SPEAKER_2 (2:40)

It is a massive sample. Every statistic we discussed today represents a group average derived from that specific aggregate data.

SPEAKER_1 (2:48)

Right. Group averages.

SPEAKER_2 (2:49)

Exactly.

We are looking at broad systemic trends in human behavior. We are absolutely not making deterministic claims about how any individual man or woman behaves in the workplace.

SPEAKER_1 (3:00)

Because everyone is different.

SPEAKER_2 (3:01)

Exactly. The value here is in understanding the broader environmental dynamics that you are operating in every single day.

SPEAKER_1 (3:08)

The methodology they used to capture those dynamics is just fascinating to me. I would so clever. It really is. The Harvard researchers wanted to test how familiar people are with the current AI landscape.

They presented these 5,400 respondents with a list of names. They simply asked if they recognized them as figures in the AI space.

SPEAKER_2 (3:27)

Just a simple yes or no. Do you know this person?

SPEAKER_1 (3:30)

Right. The list included highly visible real executives like Sam Altman, the CEO of OpenAI. He was on there. Sure. Mira Morati, the former CTO. She was on there too.

But the researchers seeded the list with a trap.

SPEAKER_2 (3:44)

The phantom variables. Yes.

SPEAKER_1 (3:46)

They inserted two names that were entirely fabricated, Susan Milner and Craig Dalton.

SPEAKER_2 (3:51)

I love this part.

SPEAKER_1 (3:52)

These individuals do not exist. They have no LinkedIn profiles. They have never committed code. They are totally made up.

SPEAKER_2 (3:58)

It is such a highly effective mechanism for behavioral research.

SPEAKER_1 (4:02)

How so?

SPEAKER_2 (4:03)

Well, by introducing phantom variables, the researchers completely remove actual factual knowledge from the equation.

SPEAKER_1 (4:10)

Because there is no knowledge to have.

SPEAKER_2 (4:11)

Exactly. The question is no longer about who actually knows the industry. It isolates the raw psychological impulse to claim familiarity when literally none exists.