Logo
Log in
Subscribe
Logo
Jo McKinney: Women Underrate, Men Overrate

Oct 15, 2026

Jo McKinney: Women Underrate, Men Overrate

Jo McKinney: Women Underrate, Men Overrate
00:00
44:46

Corrix scored 885 sessions of people working with AI. Women outperformed men and rated themselves lower. Years with AI, a tech background and company AI training didn't predict who did well. Jo McKinney on who should design the AI workplace.

Jo McKinney is a former agency CEO with more than 30 years of experience working with clients including P&G, Coca-Cola and Pfizer. Now she's co-founder and president of Corrix, which measures how people actually work with AI. At every turn in a conversation, Corrix compares what the person did with what two AIs would have produced talking to each other, so it can see what the human added. Logan saw Jo's keynote at the AI-Powered Women conference at MIT and knew within about 11 minutes that she wanted her on the show.

The finding that got the big laugh in that room: women outperformed men on a lot of measures, but underrated their performance, and men overrated theirs. Nobody's self-rating predicted anything. Neither did years of using AI, a technical background or how often you use it. What separated good work from bad was behavior. Women challenged the AI more often and kept challenging longer, while men tended to accept an answer sooner. Jo's message to anyone who thinks they're late to the party: "You probably already have what it takes to be the right kind of machine collaborator." The biggest boost to the work came from people bringing what only they know.

Logan confesses that her version of pushback is cursing at Claude, right after Mara threatens to say "double click" forever just to annoy her. Jo explains how to challenge AI when you aren't the expert: ask for sources, make it critique its own work, and play two models against each other. Then workslop, the eight-page AI report nobody asked for, and Jo's most radical idea. You shouldn't be allowed to generate with AI until you have the knowledge and judgment to evaluate what comes back.

Jo left a CEO job three years ago to get back into the work as a practitioner, and her data backs the move: individual contributors added more to their AI sessions than managers did. In Corrix's study, people who'd had company AI training scored no better than people who hadn't, which is why Jo argues this can never be a one-and-done skill. Logan explains the PhD proposal she wrote on metacognitive forcing functions.

In the drinks round, a listener in research compliance writes in. Everyone on her team uses AI to draft policies and checklists, and nobody talks about it because they're embarrassed. Jo points to research showing women are judged more harshly for using AI at work. Logan says this is the moment for practitioners to show what they've built, inside the bounds of their company's AI policy. Mara's answer: AI is a team sport, and companies have to give people time in the workday to learn it together, especially women, who have the least time to spare.

Jo's ask for listeners: take the Corrix AI collaboration assessment and see how you actually work with AI.

What needs womansplaining? Leave us a voicemail at womansplainingai.com/answering-machine or email [email protected]. Anonymous is fine. Your question could be the next drinks round.

In this episode

  • 00:00 Cold open: women outperform, and underrate themselves

  • 00:25 Welcome Jo McKinney, co-founder and president of Corrix

  • 01:32 Mara's humblebrag: she aced the Corrix assessment

  • 02:06 Jo's keynote at AI-Powered Women at MIT

  • 02:54 A new workplace, designed by a narrow set of people

  • 03:56 Measuring the quality of the work, not adoption

  • 05:02 The big laugh: women underrate, men overrate

  • 06:32 "A collective sigh of being seen"

  • 07:44 Thinking, Fast, Slow, and Artificial, and the King's College London survey

  • 08:57 How Corrix measures what the human added

  • 09:46 Nobody can self-assess their AI skill

  • 11:37 MIT Media Lab: 83% couldn't recall their own work

  • 12:02 Experience and a tech background don't predict performance

  • 12:59 Challenge the polished answer, and stay in it

  • 14:22 "I'm not techie enough"

  • 15:04 Women challenge more, and keep challenging

  • 16:13 "You probably already have what it takes"

  • 16:27 Domain expertise is the difference maker

  • 17:51 Experts should redesign their own work

  • 19:12 Double click, and cursing at Claude

  • 20:32 Frame the problem up front

  • 21:02 How to challenge AI when you're not the expert

  • 22:37 Mara's Claude argues with her Corrix results

  • 23:13 Show it what good looks like, and bring what only you know

  • 24:34 Workslop

  • 27:21 Logan: the frontier versus the slop

  • 28:02 Agents, and where humans sit in the system

  • 28:54 New workers still need to learn judgment

  • 29:12 "You shouldn't be allowed to generate"

  • 29:37 Rishad Tobaccowala on the future of work

  • 30:19 Why Jo left the CEO job to be a practitioner

  • 31:01 Individual contributors get more out of AI than managers

  • 31:35 Work doesn't reward beginners

  • 32:33 Corrix's theory of change

  • 34:11 This will never be a static skill set

  • 34:59 A desktop coach that stops you before you hit submit

  • 35:43 Logan's PhD proposal: metacognitive forcing functions

  • 36:12 Preserving human judgment

  • 36:46 Drinks round: everyone uses AI, nobody admits it

  • 37:51 Women are judged more for using AI

  • 39:16 The moment for practitioners to show their work

  • 40:01 Board of Innovation and the builder mindset

  • 41:08 Jo's weekly women's AI working group

  • 41:39 Mara: AI is a team sport

  • 42:44 Not just another software rollout

  • 43:17 Time poverty, and why this has to happen in the workday

  • 43:35 Where to find Jo and Corrix

Resources mentioned

  • Take the Corrix AI collaboration assessment

  • Corrix, "Measuring Human-AI collaboration" (885 scored sessions, September 2026)

  • Jo McKinney on LinkedIn and on TikTok (@heyjomckinney)

  • Steven D. Shaw and Gideon Nave, "Thinking, Fast, Slow, and Artificial" (Wharton)

  • King's College London's student AI survey (The Tab, May 27, 2026)

  • MIT Media Lab, "Your Brain on ChatGPT"

  • Rishad Tobaccowala, The Future Does Not Fit in the Containers of the Past (Substack)

  • Board of Innovation

Transcript

Womansplaining AI

Listen on

Apple PodcastsApple PodcastsPocket CastsPocket CastsCastroCastro

Recent episodes

Courtney Caldwell: The AI Rumspringa

Courtney Caldwell: The AI Rumspringa

Nov 12, 2026

Wanjiku Kamau: Who Do I Think I Am? (Rerun)

Wanjiku Kamau: Who Do I Think I Am? (Rerun)

Nov 5, 2026

Lizzie Hudson: Three Men and Half a Woman

Lizzie Hudson: Three Men and Half a Woman

Oct 29, 2026

My Claude Doesn't Do That

My Claude Doesn't Do That

Oct 22, 2026

I Do Not Have a Body

I Do Not Have a Body

Oct 8, 2026

Okay, Let's Slow Down

Okay, Let's Slow Down

Oct 1, 2026

Can I Hang Up?

Can I Hang Up?

Sep 24, 2026

You Can Come to the Hackathon Clueless

You Can Come to the Hackathon Clueless

Sep 17, 2026

Who Asked for This?

Who Asked for This?

Sep 10, 2026

Bring a Folding Chair

Bring a Folding Chair

Sep 4, 2026

Womansplaining AI

Women, work, and the messy middle of AI.

© 2026 Womansplaining AI.
beehiivPowered by beehiiv