← Back to articles
← Previous (older)
Why Smart People Paid $130,000 for a Steel Watch
Next (newer) →
Can AI Really Help Poor Countries Skip Ahead? A Kenyan Scientist Says: Not So Fast
technology · economics · May 11, 2026

AI Comes for the Inbox: Why Clerical Work Is the First Pink-Collar Casualty

No reader ratings yet.
Log in to rate this article
📰 Reading Passage

Jennifer Maffei, a recruiter who places executive assistants at US companies, says her inbox has become a chronicle of corporate anxiety. Workers laid off in waves of 'right-sizing' message her daily; the firms doing the laying-off — many of them household names like Procter & Gamble — increasingly ask her for staff who can already 'do x, y, z' with artificial intelligence. The pattern, she says, is bigger than any one employer. It is a structural shift in what office work means.

Clerical and administrative jobs — medical transcriptionists, receptionists, executive assistants — sit at the centre of that shift. According to the Brookings Institution, roughly 6 million US workers are most exposed to AI-driven displacement, and more than 85% of them are women. These workers tend to be older, with narrower skill sets and fewer fallback options. Job postings for administrative assistance have dropped to 5.4% below pre-Covid levels, even as women have otherwise driven recent gains in employment, adding 572,000 jobs since 2023 compared with 184,000 for men, per the Bureau of Labor Statistics.

The technology behind the squeeze is no longer hypothetical. Anthropic's Claude can schedule meetings, draft notes, and prepare documents; specialised platforms like Lindy offer end-to-end administrative support at a fraction of a human assistant's cost. Shipping giant Maersk has announced roughly 1,000 administrative cuts globally, and broader layoffs at Amazon and P&G have hit corporate support functions hard. Kelly Norton, an executive assistant earning a six-figure salary at a Las Vegas firm, says her workload has thinned to about ten hours a day from what was once a punishing schedule — and she suspects the position itself is no longer truly necessary.

But here's the catch: AI is not coming for every administrative job equally. Senior executive assistants who coordinate complex projects, plan events, and manage 'highly sensitive tasks for top-tier professionals' are, for now, relatively safe. Junior, task-heavy roles — the on-ramps to those senior jobs — are the ones disappearing. Allison Elias of the Darden School of Business warns that this matters historically: clerical work was for decades one of the few corporate ladders women could climb, and removing the bottom rungs may quietly narrow paths to management. A Harvard study last year added a second wrinkle, finding that women used AI at work at a 25% lower rate than men. The very workers most at risk are the least trained on the tools replacing them.

The International Labour Organization reported last year that female employment in higher-income countries faces nearly 9.6% exposure to AI automation, warning of 'limited opportunities to adapt' without retraining or role redesign. Jeff Strohl of Georgetown's Center on Education and the Workforce argues that occupations resilient to AI tend to bundle many different skills together, so that automating one task doesn't kill the whole job — and clerical work, unfortunately, often fails that test. Some employers are responding: a few are routing displaced admins into project-management or HR roles, others are letting workers like Maffei's former legal assistants return to paralegal pathways. Most, though, are not. Maffei's advice to clients, regardless of the path they choose, is blunt: focus on the things that still need a human.

📎 Download Original ⬇ Download Analysis PDF

📖 Explanation

The first wave of AI layoffs isn't hitting coders or radiologists — it's hitting the executive assistants, schedulers, and receptionists who keep offices running. And about 85% of them are women.

📖 What's Going On?

Generative AI tools like Anthropic's Claude and specialised platforms such as Lindy can now schedule meetings, draft emails, take notes, and prepare documents — the bread-and-butter of administrative work. Companies including Maersk, Procter & Gamble, and Amazon have already cut thousands of back-office roles as they invest in AI.

The Brookings Institution estimates roughly 6 million US workers are most exposed to AI-driven displacement, and more than 85% of them are women. Job postings for administrative assistance have fallen to 5.4% below pre-Covid levels, even as overall female labour force participation has surged.

🎯 How To Think About It

This isn't the standard story of 'robots take factory jobs.' It's something subtler: a technology that excels at one specific kind of cognitive labour — synthesising information, handling routine communication — landing squarely on a workforce that is older, more female, and historically under-trained on the tools replacing them.

💡 Key Things To Know

🌟 Why It Matters

If you're 16 and thinking about career paths, this story rewrites the old playbook. 'Get a stable office job and work your way up' assumed the bottom rungs of the ladder would still exist. They may not. Allison Elias of Darden Business School warns that clerical jobs historically functioned as launchpads — and without them, fewer women may cross into management at all. The flip side: the humans who survive are those who can do what AI can't — judgment, relationships, anticipating what should happen next.

🔮 The Bigger Picture

Every major technological shift creates winners and losers, but the losers are rarely random — they're the people closest to the tasks being automated. Watch for two second-order effects: first, pressure on governments and companies to fund retraining (Brookings and others are pushing for this); second, a possible reversal of decades of female labour-force gains if reskilling lags. The ILO has already flagged that nearly triple the share of women's employment in high-income countries faces 'limited opportunities to adapt' without retraining.

📚 Key Terms Glossary

Generative AI
AI systems (like ChatGPT, Claude, or Gemini) that produce new text, images, or code by predicting patterns from huge training datasets — as opposed to AI that only classifies or recommends.
Clerical work
Office-support jobs involving routine paperwork, scheduling, correspondence, and record-keeping — historically a major employer of women without four-year degrees.
Brookings Institution
A century-old US think tank that researches economics and public policy; its labour-market reports are widely cited by policymakers.
Labour force participation
The share of working-age adults who are either employed or actively looking for work. A key indicator of how engaged a population is with the paid economy.
Bureau of Labor Statistics (BLS)
The US federal agency that produces official employment and wage data, including the monthly jobs report.
International Labour Organization (ILO)
A United Nations agency that sets global labour standards and studies workforce trends across member countries.
Median pay
The middle value in a list of wages — half of workers earn more, half earn less. Often more meaningful than 'average' pay, which gets skewed by very high earners.
Reskilling
Training workers in new skills so they can move into different roles, typically when their old role is disappearing due to technology or restructuring.

✏️ Reading Comprehension Quiz

Tip: log in or create a free account to save your score, earn badges, and appear on the leaderboard. Otherwise the quiz works fine without an account.
Question 1
The passage most directly argues that AI's impact on the workforce is:
Question 2
Which choice best states the central idea of the passage?
Question 3
According to the passage, administrative job postings have fallen below pre-Covid levels by approximately:
Question 4
As used in the passage, the word "resilience" most nearly means:
Question 5
As used in the passage, the phrase "sharp end" most nearly means:
Question 6
Which statement about senior versus junior administrative roles can most reasonably be inferred from the passage?
Question 7
The passage suggests that one reason women are particularly vulnerable to AI displacement is that:
Question 8
The author's tone throughout the passage is best described as:
Question 9
It can most reasonably be inferred from the passage that, without intervention, AI adoption could:
Question 10
Which choice provides the best evidence for the answer to the previous question?
← Previous (older)
Why Smart People Paid $130,000 for a Steel Watch
Next (newer) →
Can AI Really Help Poor Countries Skip Ahead? A Kenyan Scientist Says: Not So Fast