--- title: "The 6.1 Million Workers AI Is Coming for First Are 86…" canonical: "https://www.metaintro.com/blog/ai-clerical-jobs-women-brookings-2026" language: "en" author: "drashtigarach" published: "2026-05-11T09:58:01.000Z" modified: "2026-05-11T09:58:01.000Z" --- [Back to Blog](/blog) [AI](/blog/tag/ai)[News](/blog/tag/news)[Information](/blog/tag/information) # The 6.1 Million Workers AI Is Coming for First Are 86 Percent Women — Inside Brookings' Sharpest Warning Yet Brookings finds 6.1M clerical workers face high AI exposure and low adaptive capacity. 86% are women. The occupation list, geography and reskilling map. [![Drashti Garach](https://cdn.metaintro.com/rs:fill:40:40/q:72/plain/images/5719d740-e510-42bc-8017-e040d145f35f_1766029465094.png)Drashti Garach @DrashtiGarach](/blog/author/drashtigarach) [May 11, 2026](/blog/archive/2026/05)12 min read ![The 6.1 Million Workers AI Is Coming for First Are 86 Percent Women — Inside Brookings' Sharpest Warning Yet](https://cdn.metaintro.com/rs:fill:1200:675/q:78/plain/images/FILES.png) [https://x.com/intent/tweet?text=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026](https://x.com/intent/tweet?text=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026)[http://www.facebook.com/sharer.php?u=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026](http://www.facebook.com/sharer.php?u=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026)[https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026&title=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026&title=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet)[mailto:?subject=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&body=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026](mailto:?subject=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&body=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026) Brookings dropped a research note in early 2026 that the wider press is only catching up to now: of roughly 37.1 million U.S. workers in the top quartile of generative-AI exposure, about 6.1 million sit in the worst quadrant — exposure high, adaptive capacity low. Roughly 86 percent of them are women, concentrated in clerical and administrative occupations. That is not a forecast of who might be displaced; it is a map of who is least equipped to land somewhere else when displacement happens. The [Cloudflare layoffs](https://www.metaintro.com/blog/cloudflare-layoffs-2026) announced May 7 — 1100 support roles cut as AI usage inside the company rose 600 percent — show the timeline is no longer hypothetical. ## Who exactly is in the Brookings high-risk group? The Brookings team behind ["Measuring US workers' capacity to adapt to AI-driven job displacement"](https://www.brookings.edu/articles/measuring-us-workers-capacity-to-adapt-to-ai-driven-job-displacement/) did not stop at a top-line exposure score. They overlaid AI-exposure rankings against an adaptive-capacity index built from savings, age, local labor-market depth and skill transferability. The intersection — high exposure plus low capacity — is where 6.1 million workers live, and the occupations doing most of the lifting are not the ones that dominate Substack debates about AI. The list named in the Brookings analysis includes secretaries and administrative assistants (excluding legal, medical and executive variants), payroll and timekeeping clerks, court, municipal and license clerks, property appraisers and assessors, tax examiners and revenue agents, eligibility interviewers for government programs, medical transcriptionists, receptionists and executive assistants. These are exactly the [administrative roles](https://www.metaintro.com/blog/administrative-interview-questions) the U.S. Bureau of Labor Statistics has tracked as core office-support employment for decades. The Bureau's own [2024-34 occupational projections](https://www.bls.gov/opub/mlr/2026/article/industry-and-occupational-employment-projections-overview.htm) already flagged secretaries and administrative assistants (except legal, medical and executive) for decline on top of the AI overlay. Why 86 percent women? Because those occupations were already 80-94 percent female before generative AI shipped. Brookings is not making a prediction about gender — it is reading existing demographic concentration into a new exposure curve. The result is the sharpest demographic disparity in any major AI workforce study to date, and the reason secondary coverage from outlets like [Inc.](https://www.inc.com/kit-eaton/ai-poses-a-greater-job-threat-to-women-than-men-new-data-shows/91342399) led with the demographic angle rather than the geography. ## What does "low adaptive capacity" actually mean in this report? Adaptive capacity in the Brookings framework is built from four inputs: liquid savings to weather a job search, age (older workers have shorter runway to retrain and shorter post-retraining earnings tail), depth of the local labor market in adjacent occupations and breadth of transferable skills. A clerical worker in her early 50s, in a smaller metro, with under three months of savings and a resume of mostly office-administration tasks scores near the floor on every input. That is not a character verdict — it is what the labor market looks like for women who built careers in the back office of mid-sized employers in the 1990s and 2000s. Compared with the broader pool of 37.1 million highly exposed workers, about 26.5 million land above the median on adaptive capacity. Their displacement story will be ugly but recoverable. The 6.1 million who do not — they are who the policy debate is failing to name out loud. Researcher reactions in [related Brookings work](https://www.brookings.edu/articles/generative-ai-the-american-worker-and-the-future-of-work/) on the geography of AI's workforce impacts back this up: the cushion shrinks fast outside the top 25 metros. ## Where do these workers live? Geography is the second twist most coverage misses. The Brookings team found that the share of workers in the high-exposure / low-capacity quadrant is highest in college towns and state capitals, especially across the Mountain West and Midwest — places like Lincoln, Topeka, Cheyenne, Boise and Madison rather than the coastal metros driving AI venture capital. State governments and public universities run on the exact administrative spine the technology is now consuming. That matters for two reasons. First, when a public-sector clerical role evaporates in a mid-sized state capital, the nearest comparable role might be hours away, not the next office park. Second, these labor markets have thinner gig and contract substitutes than New York or San Francisco, so the bridge that some [women in metro areas](https://www.metaintro.com/blog/women-contract-workers-metro-areas) use during transitions is not actually there. The geography also stress-tests one of the most common reassurances in AI labor coverage: that workers can simply relocate. In practice, the Brookings map shows the high-exposure / low-capacity quadrant clustered in metros where housing markets are tighter than coastal stereotypes suggest, where two-earner households have spousal jobs anchored to the same employer or university and where school systems and elder-care arrangements are already calibrated to local life. Moving is not a free variable. The places taking the first wave of clerical AI displacement are also the places where the financial and social cost of leaving is highest. That is part of why the adaptive-capacity index weights local labor-market depth so aggressively in the first place — it is encoding the fact that an open job 600 miles away is, for most clerical workers in their late 40s and 50s, not really an open job. ## Are Cloudflare and Salesforce already proving the thesis? Cloudflare CEO Matthew Prince told investors May 7 that the company cut 1100 employees — roughly 20 percent of headcount — even as it posted record revenue of $639.8 million for Q1 2026, up 34 percent year-over-year. He drew a hard line between "people who build the product, people who sell the product, and people who support the people who build and sell the product." The third group is the one being replaced. The Cloudflare layoff math — record revenue, 600 percent rise in internal AI use, support roles gone first — is the Brookings model running in real time at one company. Salesforce ran the same play earlier this year, with a [February reorg](https://www.metaintro.com/blog/salesforce-layoffs-february-2026-agentforce-ai) that cut customer support staffing as Agentforce absorbed inbound tickets, and [CEO Marc Benioff has been explicit](https://www.metaintro.com/blog/salesforce-ceo-ai-agents-gig-workers) about agents replacing layers of work. Both companies are doing the version of AI deployment that Brookings warned would hit clerical and support roles first — and both are doing it on rising revenue, which neutralizes the usual "they had to cut" narrative. The pattern extends beyond these two names. IBM has spent 2026 publicly tying back-office automation to internal AI agent rollouts, and several large telecoms have quietly reorganized customer-care functions around AI triage layers that route only escalations to humans. The common thread is that each company is keeping the workers AI cannot yet replace — product engineering, enterprise sales, executive judgment — and removing the workers AI can. That is not a hypothetical alignment with the Brookings analysis; it is the same occupational map turned into an org chart. For job seekers, the signal is that the company-by-company timing is now visible in earnings calls and 10-Qs rather than hidden in research papers. Watching which firms post record revenue while announcing support-side cuts is the cleanest leading indicator of which clerical labor markets will tighten in the next two quarters. ## Which reskilling paths actually move women into safer roles? Most reskilling-program optimism is downstream of [corporate programs that are failing the workers they target](https://www.metaintro.com/blog/corporate-reskilling-programs-failing-workers-future-proof-career-2026). The Brookings data is more useful as a transition map than as a doom forecast — it points to adjacent occupations where the same core skills (judgment under ambiguous rules, customer empathy, structured documentation) carry over but AI exposure drops. Three transitions are doing real work in the data. Clerical-to-allied-health: medical office and revenue-cycle specialists move into patient navigation, care coordination and certified medical assisting. Clerical-to-skilled-trades-adjacent: payroll and timekeeping experience translates into construction or facilities project administration. Clerical-to-public-safety-administration: court clerks and eligibility interviewers move into investigator, compliance and regulatory roles that still demand human judgment. The federal [Department of Labor's $243M AI apprenticeship initiative](https://www.metaintro.com/blog/dol-ai-apprenticeship-initiative-243-million-reskilling-career-path-2026) is funding several of these bridges, though uptake is uneven. The transitions that do not work in the data: clerical-to-data-analyst pipelines (the entry rung is already AI-saturated), clerical-to-software-support (Cloudflare just told you why) and clerical-to-prompt-engineer (a job category that is collapsing back into ordinary product work). Reskilling has to be honest about which doors are still open, which is the spirit of our [AI reskilling gap analysis](https://www.metaintro.com/blog/ai-reskilling-gap-companies-failing-workers-what-job-seekers-should-do-2026) for job seekers. ## What can policy and employers actually do? The Brookings authors are careful to argue this is not a story of inevitable displacement. Their framing is closer to: if you know who is most exposed and least equipped to adapt, the policy levers are visible. State-level [90-day AI-displacement notice laws](https://www.metaintro.com/blog/minnesota-ai-displacement-90-day-notice-bill-worker-reskilling-rights-2026), like the bill moving through Minnesota, give workers planning runway. Public-sector employers — the ones running on the exact clerical spine Brookings flagged — can pace AI rollouts against internal mobility programs rather than headcount cuts. For employers, the most defensible pattern is what the Brookings note implicitly endorses: redeploy clerical staff into roles where AI handles the routine and humans handle escalation, regulation and edge cases. That requires job redesign budgets most companies are not funding. The alternative is the Cloudflare pattern — cut, redirect savings to AI capex, accept the reputational drag. Brookings is telling policymakers and employers exactly which workforce will absorb that drag if the redeploy lever stays unfunded. There is also a state-level fiscal story most policy coverage is skipping. State governments and public universities are themselves among the largest employers of the high-exposure clerical workforce Brookings flagged. That means the same state legislatures debating AI displacement notice laws are also the entities making procurement decisions about AI tools that will displace their own administrative staff. The cleanest interventions tie the two together: any state agency adopting generative AI for administrative work commits up front to redeployment funding equal to a percentage of the productivity savings, with priority for affected workers. A few state CIO offices are already drafting that language. The federal lever is narrower — DOL apprenticeship dollars and Workforce Innovation and Opportunity Act funds can subsidize the bridge programs — but state procurement is where the pace of clerical displacement will actually be set over the next 18 months. The dollars are there if state procurement teams decide the cost of retraining is cheaper than the cost of community-level layoff fallout — and on the math Brookings is laying out, in most state capitals on this list, it clearly is. ## What can individual workers do this quarter? If you are in one of the named occupations, the practical playbook is shorter than the policy debate. First, run the [AI-displacement survival checklist](https://www.metaintro.com/blog/ai-displacement-survival-playbook-2026) — pull a current resume, identify three adjacent occupations using O\*NET skill overlap and price out a 12-week certification in the strongest match. Second, build a 90-day cash buffer; Brookings's adaptive-capacity index weights savings heavily for a reason. Third, document the judgment calls, escalations and exceptions you handle that AI agents currently mangle — that is the case for your redeployment, not your replacement. And resist the framing that this is a personal failure of skill. The Brookings analysis is structural. The 86 percent figure exists because mid-century career advice told women to take stable office work and the labor market priced that stability fairly until generative AI changed what "stable" means. That is the context the [40 percent of workers fearing AI job loss](https://www.metaintro.com/blog/40-percent-workers-fear-losing-job-to-ai-2026) in recent surveys are reacting to, even if they cannot name the Brookings paper. Coverage threading the same gender lens, from the [Irish Times](https://www.irishtimes.com/business/2026/05/11/women-at-the-sharp-end-as-ai-takes-over-administrative-roles/) to ongoing tracking of [women's AI job-loss risk](https://www.metaintro.com/blog/women-risk-ai-job-losses), is converging on the same conclusion. ## People also asked **Q: How many U.S. workers face the highest AI displacement risk according to Brookings?** A: Brookings identified about 6.1 million U.S. workers who are both highly exposed to generative AI and have low capacity to adapt to displacement. That is a subset of roughly 37.1 million workers in the top quartile of AI exposure overall. **Q: Why are women 86 percent of the highest-risk group?** A: Because the occupations that combine high AI exposure with low adaptive capacity — secretaries, administrative assistants, payroll clerks, receptionists, court clerks, medical transcriptionists — were already 80-94 percent female before generative AI arrived. Brookings is reading existing demographic concentration into a new exposure curve. **Q: Is the Cloudflare layoff really an example of this trend?** A: Yes. Cloudflare cut 1100 support-function jobs on May 7 while posting record revenue, citing a 600 percent rise in internal AI use. The CEO explicitly drew the line between product, sales and support roles, with support being replaced. That is the Brookings model deployed at one company. Future-proof your career. [Metaintro](https://www.metaintro.com) matches your background to roles built for the AI workforce era. ### Share this article [https://x.com/intent/tweet?text=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026](https://x.com/intent/tweet?text=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026)[http://www.facebook.com/sharer.php?u=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026](http://www.facebook.com/sharer.php?u=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026)[https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026&title=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026&title=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet)[mailto:?subject=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&body=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026](mailto:?subject=The%206.1%20Million%20Workers%20AI%20Is%20Coming%20for%20First%20Are%2086%20Percent%20Women%20%E2%80%94%20Inside%20Brookings'%20Sharpest%20Warning%20Yet&body=https%3A%2F%2Fwww.metaintro.com%2Fblog%2Fai-clerical-jobs-women-brookings-2026) For job seekers ## The right role, found for you. Built for people who’d rather be good at their job than good at job hunting. [Get Started Free](/signup)[Search matching jobs](/jobs/search) - MatchCompare live roles against your current evidence. - PositionTurn proof projects into role-specific applications. - ImproveUse market feedback to keep the skill plan current. [Return to navigation](#main-navigation)