AI-washing: when "AI" is the reason in the press release

On 10 November, Qantas told staff that 850 jobs in its head office would go. Its chief executive explained the change with AI: “Leading organisations around the world are well progressed in using technology and AI to completely transform how they work.” Then came the detail: most of the work moves to Accenture, and most of Accenture’s service will be delivered from India. That is not a machine replacing a person. It is a person in one country replacing a person in another, with AI in the press release. Critics have a name for this: AI-washing.

Contents · 9 sections

TL;DR #

• Qantas framed 850 job cuts around AI, but the work mostly moves to people in India.
• Australia’s parliament is asking banks which cuts were AI and which were offshoring; the union estimates more than 10,000 bank jobs have gone in two years.
• A vendor survey says about one in five US respondents were replaced by AI and then asked back. It is a vendor survey, so treat it with care.
• My news engine found that AI-framed layoff stories reach dozens of outlets from only 2 to 5 original voices, mostly through syndication.
• Seven practical tests help tell a real AI cut from a label.

My earlier post on the 2023–2026 wave covered the background research, so I only summarise it here. All facts are as of 13 November 2026.

The background in one paragraph #

The case that AI is over-credited for layoffs was already strong. Oxford Economics found AI was cited for about 4.5% of US job cuts in 2025 and that productivity growth slowed, the opposite of what mass substitution would produce; one of its economists suspected firms were trying to “dress up layoffs as a good news story”, according to Fortune. Klarna replaced customer-service staff with AI and then started hiring humans again after its chief executive told Bloomberg that cost had been “a too predominant evaluation factor”. November added new cases, and one new kind of evidence.

Qantas: offshoring with an AI label #

The Qantas announcement was the most independently covered layoff story of the month in my data: 156 outlets, 132 independent voices, 18 languages (LSN News). What is documented:

• About 850 roles are affected “over the coming years”. The first phase covers finance, procurement, human resources and marketing, by mid-2027.
• The work goes to Accenture under a technology partnership. According to 1News and Reuters, most Accenture services will be delivered in India, with some roles staying in Australia.
• The company framed it as modernising and becoming more efficient with technology and AI.

The Australian Services Union’s NSW/ACT secretary did not mention AI at all: “This is skills offshore at our national airline and if they can get away with it, well, anyone can.” The union says Qantas had earlier assured it there was no offshoring plan; Qantas denies giving that assurance.

The release does not say how many of the 850 roles will be done by AI and how many by people in India. That missing number is the whole question.

In my library, Daniel H. Pink’s A Whole New Mind names three forces reshaping white-collar work: Abundance, Asia and Automation. Qantas is a case of Asia wearing Automation’s clothes. Offshoring back-office work is a twenty-year-old strategy. What is new is the label.

Australia’s banks: an inquiry into the label #

The Australian Associated Press reported on 11 November that bank bosses face a parliamentary inquiry over AI and offshore job cuts. The figures in that story:

• The big four banks have cut, or plan to cut, about 6,000 jobs since September 2025.
• More than 10,000 Australian banking jobs have gone in two years, according to the Finance Sector Union’s estimate, many offshored or replaced by AI.
• ANZ cut 3,500 staff in the year to September, reported on the same day it posted a 23% rise in statutory net profit to A$7.27 billion.

The inquiry’s point is the one that matters here: when “AI” and “offshore” sit in the same sentence, the public deserves to know which is which. As of 13 November no findings have been published, so this is documented as an inquiry, not as a verdict.

“Replaced by AI, then asked back” #

A survey by SHL, a company that sells workforce-assessment tools, reported in early November that 19.1% of US respondents said they had personally been replaced by AI and later asked to return. Among 18- to 24-year-olds the figure was 24%. It covered 2,036 workers in the US and UK in June and July 2026, and Fortune and others carried it as “AI rebounding”.

I grade this as a vendor survey. It is self-reported, from a company with a commercial interest in the topic, and “replaced by AI” means whatever each respondent understood it to mean. I would not quote the 19% as a fact about the labour market. What the survey does show is that “we replaced them with AI” and “we need them back” now sit in the same news cycle, a sign the label is wearing thin.

A counter-example comes from the employer side. When Microsoft cut about 4,800 roles in July, its chief people officer wrote: “the roles eliminated today are not being replaced by AI.” A company saying that plainly is rare, and worth noting.

Many outlets, few voices #

LSN News, my news engine, groups articles into stories and counts independent voices: outlets that wrote their own copy, as opposed to outlets that republished the same text from a wire or a sister paper. That distinction matters a lot for the AI-layoff story.

The AI-framed layoff stories that looked big were mostly syndication:

• “Bank bosses face inquiry over AI and offshore job cuts” reached 52 outlets with only 2 independent voices. All 52 belong to one regional newspaper group, Australian Community Media (LSN News).
• A story on an S&P Global Ratings report warning that the US-China AI race could bring financial setbacks and mass layoffs over three years reached 25 outlets with 2 independent voices; 23 of them sit in one aggregator network (LSN News).
• An AP explainer on what to consider if a company that made AI-related job cuts wants you back reached 45 outlets with 5 independent voices, about three-quarters of them in one US newspaper group (LSN News).

Compare those with stories about named employers and hard numbers: Qantas had 132 independent voices, Boeing 99 and Vodafone 78. Meanwhile an AI-first corporate cut got little independent attention: Canva’s AI-driven restructuring reached 7 outlets (LSN News).

Two more numbers from the same week. About 9% of distinct layoff headlines mentioned AI or automation, and most of those were commentary, not new cuts. On 10 November, 69 English-language articles mentioning AI and layoffs collapsed to 7 distinct headlines once duplicates were removed.

So the AI-layoff narrative, in both directions (AI is taking jobs, and firms are rehiring after AI), reaches dozens of outlets from 2 to 5 original voices. The mechanism is ordinary syndication: one group republishing one piece. My data shows no coordination beyond that, and it cannot tell republishing deals from editorial intent. But if you judge how widely a framing is held by how many sites carry it, you will overestimate it.

The limit: LSN News has history only from 7 November 2026. This is one week, and it says nothing about trends.

Checklist: seven tests for an “AI” layoff #

A cut can be partly about AI and partly about something else. These tests won’t give a verdict, but they show which questions to ask.

1. Where does the work go? If an outsourcer or another country is named, the main story is sourcing, not software. Qantas fails this test.
2. Is a named system doing the work, now? “AI will transform how we work” is a forecast. “This tool now handles 70% of these tickets” is a claim you can check. Watch for “over the coming years”.
3. Which roles shrink? If cuts hit the specific tasks a deployed tool does, AI is plausible. A flat percentage across departments is a budget.
4. What do the profits say? Cutting while profits rise is not proof of AI-washing, but it means the cut was a choice. ANZ cut 3,500 jobs in a year its profit rose 23%.
5. What does the company tell people who can hold it to account? Compare the press release with what is said to a works council, union, regulator or inquiry. Disclosures under consultation law tend to be more specific than investor slides.
6. Is it hiring the same skills back? Rehiring, even through contractors, means the work did not go to a machine. Klarna is the classic case.
7. How many voices carry the framing? Before treating “AI is replacing X” as consensus, check whether it is fifty outlets or one wire repeated fifty times.

Frequently asked questions #

What is AI-washing? #

Presenting job cuts, or other decisions, as driven by AI when the main drivers are something else, such as cost cutting, offshoring or weak demand. The term is used by critics and unions; companies rarely accept it.

Was the Qantas cut AI-washing? #

The documented facts are that Qantas framed it around AI and that most of the work goes to Accenture, largely delivered from India. The company has not said how many roles AI will do. That makes the AI framing unproven, not disproven.

Is the “replaced by AI then asked back” figure reliable? #

It comes from a vendor survey of 2,036 workers, self-reported. It is a signal that rehiring after AI cuts happens, not a measurement of how often.

Why does the AI-layoff story seem everywhere? #

Partly because syndication multiplies a few original pieces across dozens of sites. In my data, several AI-framed layoff stories reached 25 to 52 outlets from just 2 to 5 independent voices.

The lesson #

“AI” in a layoff press release is a claim, not a fact. The label rose when markets rewarded it and is fading now that it draws inquiries, union pushback and rehiring stories; Challenger’s October count had AI cited for under 3% of US cuts. The real early effect of AI on jobs is quieter, in junior hiring. The next time AI is the headline reason for a cut, ask where the work goes and how many voices are telling the story.