Will Women Physicians Face an AI Gap? Female doctors already spent 40 minutes more per day in the EHR

AI was supposed to make medicine lighter. For many women physicians, it risks making an already unequal workload even heavier. That’s the warning from Dr. Rochelle Walensky in a recent JAMA+ AI Conversations episode with Dr. Roy Perlis, discussing her viewpoint with Dr. Nicole McCann, Women, Medicine, and AI.

The productivity paradox

AI is arriving after 20-30 years of exploding administrative work without a matching growth in physicians.

In 2006, an ED chart had 5 notes to review. In 2022, the same chart had 359 notes.

20% of charts are now longer than Moby Dick

EHRs, quality reporting, billing, prior auths, portal messages — all grew. The number of doctors doing that work did not. AI should help summarize charts and manage inboxes, but tools that promise efficiency don’t always deliver it.

Why the burden falls more on women

Women physicians, on average, spend more time in the EHR and on documentation:

40+ minutes more per day in the EHR than men

30 minutes more per day on notes

50% more patient messages than men, even after controlling for panel size and number of appointments

Over a career, that compounds into less time for care, research, leadership, and rest — and more burnout.

Walensky is careful not to frame this as “women do this, men do that.” But the differences are consistent, and they coexist with another finding: in a large Medicare study, patients of female physicians had lower 30-day mortality and readmission rates. We don’t fully know what drives that, but it suggests women may be spending time differently, in ways that matter.

The AI gap

Women are using AI less, even though they have more to gain from it because their administrative burden is higher.

Globally, AI adoption is 47.8% for men vs. 39.3% for women. In medicine, the gap matters more because the specialties with the most women — family medicine, internal medicine, pediatrics, infectious disease — are also the ones with the most potential for AI augmentation. They’re cognitive, message-heavy, documentation-heavy fields.

Walensky’s point is augmentation, not replacement. AI is good at knowledge and board exams. It’s not good at holding uncertainty, knowing when to ask the right history question, or sitting with a family at the bedside.

But augmentation only helps the people who actually get to use it.

Avoiding AI won’t protect women physicians. The fix is to name the unequal burden, design AI for the work women actually do, make engagement safe and supported, and include women in building and evaluating the tools.

If we don’t, AI won’t fix inequity. It will encode it — faster, at scale.

Sources

Original podcast:

Will Women Physicians Face an AI Gap? — JAMA+ AI Conversations, JAMA / AMA Ed Hub, Sept. 28, 2026: https://edhub.ama-assn.org/jn-learning/audio-player/19095916?resultClick=1&bypassSolrId=M_19095916

Original viewpoint:

Walensky RP, McCann NC. Women, Medicine, and AI. JAMA Internal Medicine. Published online Sept. 28, 2026. doi:10.1001/jamainternmed.2026.4880: https://doi.org/10.1001/jamainternmed.2026.4880

Studies discussed:

Chart growth — “Call Me Dr. Ishmael”: https://academic.oup.com/jamiaopen/article/7/2/ooae039/7679516

EHR time by gender: https://www.ama-assn.org/practice-management/physician-health/office-based-physicians-gender-inequity-pops-ehr

Patient outcomes by physician sex: https://pqilab.dgsom.ucla.edu/treatment-from-female-doctors-leads-to-lower-mortality

Portal message surge: https://www.ajmc.com/view/patient-portal-messages-outpace-office-visits

Global AI gender gap: https://www.hbs.edu/ris/Publication%20Files/25-023_be8fb517-3dd5-40aa-97f9-4e42e1c8e6ff.pdf