Fei Zhaojun's boss asked him a simple question this month: could AI soon replace human coders? Two weeks later, Fei and 160 of his colleagues in Beijing were laid off. It reads like exactly the story everyone expects from 2026, and in China's case, it largely is. But a new Harvard Business Review analysis, citing Goldman Sachs research, argues that most of the layoffs getting blamed on AI elsewhere aren't actually AI's doing at all.
Employer branding works better when people share what they know. EBClub brings practitioners together to exchange ideas, compare experiences, and learn from the people doing the work.
This week's Paper Cut is about the gap between the layoffs AI is actually causing and the ones getting blamed on it anyway, plus what employers are quietly admitting about the rest of the deal with their people.
This Week’s Paper Cut's
Watch or listen here ↯
We keep tracking to a minimum. The video is hosted on YouTube and the podcast is hosted on PodBean, which may place cookies that measure viewing and personalize content. We only load the player after you consent to these cookies. You can update your cookie settings any time.
"AI-washing" the layoffs: the numbers don't back up the story
Harvard Business Review argued this week that many layoff announcements are better understood as "AI-washing," ordinary restructuring dressed in AI language to reassure investors that leadership has a grip on the coming technological change. The analysis cites Goldman Sachs' own estimate that AI reduced monthly US payroll growth by only around 16,000 jobs over the past year, adding roughly 0.1 percentage points to the unemployment rate, a real number, but a small one next to the volume of companies citing AI as the reason for cuts.
That gap showed up in this week's smaller stories too. Apple confirmed 147 layoffs, largely engineers at its Cupertino headquarters, without offering much explanation either way. Microsoft, meanwhile, is trying to reassure its own staff that its data centre build-out is a net positive for society, Bloomberg reported, an unusual thing to have to say internally unless people are already asking. And TechRadar reported that some Microsoft employees have been spending up to $28,000 in AI compute tokens in a single 28-day period, a trend dubbed "tokenmaxxing," where usage is inflated to look productive on internal dashboards rather than to actually get more done.
Employer Brand POV
Test the AI-blamed number against your own before you repeat it externally. Goldman Sachs puts AI's real employment drag at roughly 16,000 jobs a month nationally, nowhere close to the scale most "AI transformation" announcements imply. If your own redundancy comms lean on AI as the explanation, check whether that's actually true of the roles being cut, or whether it's the easier story to tell than the real one underneath it.
What losing a job to AI actually looks like, up close
Fei's story, reported by Fortune and the Associated Press, is one version of it: a mid-level coder in Beijing, told by his boss that AI was close enough to doing his job, laid off with 160 colleagues two weeks later. He's now making vlog-style videos during a career break rather than actively job hunting. Elsewhere in China, a part-time translator told Fortune his industry's pay has been cut by more than half in a few years as AI-assisted translation tools have spread, even as the work itself hasn't disappeared. The share of Chinese industrial firms using AI models and agents jumped to 47.5% last year from 9.6% in 2024, according to the market intelligence firm IDC, and a separate International Labour Organization report found women face higher risk of AI-related job loss than men, concentrated in roles like electronics assembly that are more exposed to automation.
The Washington Post reported a Western echo of the same pressure this week: Volkswagen's chief executive warned German workers that sweeping layoffs are coming, without yet putting a number on it. Different economy, different specifics, same shape.
Employer Brand POV
Warn your workforce specifically about which roles are actually exposed, rather than letting a general AI anxiety sit unaddressed. Fei's account and the wage data around it show the damage landing unevenly, on mid-level coders and translators, not evenly across a company. If you can't yet say which roles in your organisation look more like Fei's than not, that uncertainty is worse for morale than a hard answer would be.
Companies are rewriting what they owe their people, out loud
Netflix cofounder Reed Hastings told Semafor this week that cutting a third of Netflix's staff after the dotcom bust taught him a hard lesson: a company that calls itself a family shouldn't ever do a layoff, "you would never lay off two of your kids." Fortune's report paired that with Airbnb's Brian Chesky, who admitted his own "family" language during a 25% pandemic-era layoff round blurred a distinction he later had to walk back, and Shopify's Tobi Lütke, who told staff in 2021 that "Shopify is a team, not a family" specifically to stop the word getting in the way of accountability.
Disney offered a live example of the genre this week. An internal memo from chief people officer Sonia Coleman, offering senior leaders a Voluntary Early Retirement Offer, used the phrase "deeply personal decision" to describe a program built entirely around headcount reduction, while separately confirming involuntary cuts are already under way and will continue into next year.
Employer Brand POV
Draft the honest version of your redundancy language before a scandal or a euphemism writes it for you. Hastings, Chesky and Lütke all reached the same conclusion from the other side of a layoff: the "family" framing doesn't survive contact with the decision it's supposed to soften. Say what the change actually is, in plain terms, and save the warmer language for the parts of the relationship that are actually still true.
The entry-level market is shrinking, and the pitch to young workers hasn't caught up
Just 8,383 graduate jobs were advertised in the UK in July, according to recruitment site Adzuna, down 45% year on year and the lowest since the company started tracking the data in 2016; there were more than 55,000 graduate postings in 2017. In Australia, National Centre for Vocational Education Research figures show 52% of apprentices who started in 2020 didn't complete their training within the expected four years, with low pay, weak mentoring and long commutes cited as the main reasons, and active apprenticeship contracts fell 8.6% in the year to December 2025. A separate report from the British Standards Institution found 39% of employers have already cut entry-level roles due to AI efficiencies, with a further 43% expecting to do so this year, according to HR analysis firm Zywave.
Glassdoor's own survey data adds a generational wrinkle: only a third of Gen Z workers speak positively about AI's impact on their jobs, against 47% of Gen X, and the gender split is sharper still, 21% of Gen Z women positive versus 42% of Gen Z men. Meanwhile Gartner's 2026 survey of 10,055 employees found financial security, healthcare protection and long-term wealth creation overtaking work-life balance and paid time off as top priorities, and a separate Gartner survey of 11,838 employees found fewer than 35% report high overall wellbeing despite years of employer investment in wellbeing programmes. The Wall Street Journal reported lawmakers are now moving to ban "ghost job" postings, roles that are advertised but never filled or don't exist at all, adding a trust problem on top of a shrinking one.
Employer Brand POV
Ask what your own workforce actually prioritises before you design next year's rewards package around what they wanted three years ago. Gartner's data shows financial security overtaking work-life balance nationally, and if your benefits strategy is still built around the old assumption, you're solving a problem your people no longer rank first. The same goes for entry-level hiring: if graduate roles are this scarce and this contested, a ghost job posting or a vague AI-anxiety memo will cost you more trust than it's worth.
Related reading on EBN: Inside LinkedIn's Job Marketplace: What the "Top Picks" Debate Tells Talent Leaders About Trust
The Cut List
Goldman Sachs estimates AI has cost the US economy roughly 16,000 jobs a month, a rounding error next to the number of companies citing AI as the reason for cuts.
UK graduate vacancies have fallen 45% in a year to 8,383 roles, the lowest since tracking began in 2016, against a record number of students starting university this September.
None of this week's numbers point the same way either. Goldman Sachs says AI has cost the US economy about 16,000 jobs a month, a rounding error next to the number of companies citing AI as the reason for their cuts. Reed Hastings says the honest move is to stop calling your company a family before you do the layoff that proves it wasn't one. Disney's HR memo manages to describe a program built entirely around headcount as a "deeply personal decision." And the youngest people trying to get into any of this are applying to a graduate market that's shrunk by nearly half in a year, for jobs that increasingly might not even be real. The pattern underneath all of it: the story a company tells about a decision and the actual mechanics of that decision keep drifting further apart, and the people affected are the ones left doing the work of closing the gap.
Takeaways
- A new HBR analysis, citing Goldman Sachs, estimates AI reduced US monthly payroll growth by only around 16,000 jobs over the past year, adding roughly 0.1 percentage points to unemployment.
- A Beijing coder was laid off alongside 160 colleagues two weeks after his boss asked whether AI could replace human programmers.
- The share of Chinese industrial firms using AI models and agents jumped to 47.5% last year from 9.6% in 2024, according to IDC.
- Volkswagen's CEO warned German workers this week that sweeping layoffs are coming.
- Reed Hastings says Netflix's 2001 layoff of a third of its staff taught him that companies shouldn't call themselves families.
- Disney's chief people officer described a voluntary early retirement programme built around headcount reduction as a "deeply personal decision."
- UK graduate job vacancies fell 45% year on year to 8,383 in July, the lowest since tracking began in 2016, according to Adzuna.
- 52% of Australian apprentices who started in 2020 didn't complete their training within four years, citing low pay, weak mentoring and long commutes.
- Only a third of Gen Z workers speak positively about AI's impact on their jobs, against 47% of Gen X, according to Glassdoor.
- A Gartner survey of 10,055 employees found financial security overtaking work-life balance as a top workplace priority.
