A landmark study confirms that junior roles are being shed. Two new research papers explain what is really happening to early-career workers – and why the damage is concentrated at the bottom of the ladder, not the top
For years, economists, technologists, and executives have debated which workers artificial intelligence would most displace. The answer is becoming harder to ignore: not senior professionals, not experienced managers, but the people just starting out — the graduates, the analysts, the junior associates who were supposed to be the future of their industries. Two major reports released in recent weeks, read together, make this picture disturbingly clear.
The CFOs are telling you directly
The Oliver Wyman Forum and the New York Stock Exchange surveyed nearly 500 chief financial officers – executives whose public companies alone represent roughly 12% of global market capitalisation – about how the finance workforce will change over the next three years. The findings are unambiguous on one point: junior roles are where the pressure is being applied.
It found that 64% of CFOs expect the finance function to shift away from junior roles, while 91% anticipate either flat (61%) or lower (30%) headcount overall in finance. The traditional pyramid – wide at the base with entry-level staff, tapering to a narrow band of senior executives — is, in their own words, beginning to flatten into a middle-heavy diamond.
What is driving this? The report is clear: AI. Four in five CFOs view embedding AI in the finance function as a top three transformation priority. The use cases they are targeting – planning and forecasting, controls and fraud detection, order-to-cash, spend analytics – are precisely the areas where junior analysts have historically spent the most time. Automation absorbs the work, and the roles built around it begin to disappear.
Crucially, senior roles are not shrinking in the same way. The most common expected shift in workforce composition is toward midlevel roles (cited by 41% of CFOs), followed by a shift toward senior roles (23%). Only 13% expect a shift toward more junior roles. The pull is unmistakably upward. The base of the career ladder is under pressure; the top is not.
The research now proves it
While CFO survey data captures intentions, two significant research papers provide empirical weight to what is actually happening in the labour market right now.
A Harvard working paper by Seyed Hosseini and Guy Lichtinger, published in May 2026, analysed résumé and job-posting data covering 65 million workers across more than 280,000 US firms between 2015 and 2025. Their finding is striking: at firms that adopted generative AI, junior employment declined by approximately 9 per cent after six quarters relative to non-adopting firms. Senior employment, over the same period, showed no comparable break in trend – in fact, it continued to rise.
The mechanism matters as much as the magnitude. The decline was not driven by mass redundancies of junior staff. It was driven primarily by a sharp reduction in hiring. Companies that adopted generative AI simply stopped bringing in as many people at the bottom. The separation rate for juniors at adopting firms actually fell slightly – it is not that firms are making their existing junior workers redundant in large numbers. They are not replacing them, and not recruiting new ones.
The Harvard researchers call this "seniority-biased technological change" – a concept that inverts the familiar anxiety about AI threatening experienced workers. The jobs being eroded are not the complex, judgement-intensive roles at the top of the hierarchy. They are the entry-level, task-heavy roles at the bottom: the debugging, the document review, the data entry, the routine analysis. Those are the tasks generative AI handles best, and those are the tasks that used to define the early years of a professional career.
A parallel Stanford Digital Economy Lab study, analysing payroll records from ADP – the largest payroll software firm in the United States – found that employment for young software developers aged 22 to 25 declined nearly 20% by July 2025 from its peak in late 2022. Senior employment in the same sector showed no such decline. The pattern matches: AI exposure predicts junior employment loss, specifically in the occupations and tasks most susceptible to automation.
The gap between AI ambition and delivery – and what it means for junior staff
The Oliver Wyman report reveals an important nuance that makes the junior employment picture more complex, not less concerning. Despite the overwhelming intention to use AI in finance, actual deployment at scale remains very limited: only 8% of CFOs have deployed AI-assisted tools or autonomous agents at scale across their finance functions. Between 70 and 81% are still in the planning or piloting stages across key use cases.
This matters enormously for the junior employment question. If only 8% of finance functions have genuinely deployed AI at scale, yet 64% already expect to shed junior roles, it means much of the junior workforce contraction is driven not by automation that has already happened, but by anticipation of automation that is coming.
The Harvard researchers identified the same forward-looking dynamic: firms began reducing junior hiring shortly after the release of ChatGPT in late 2022, before large-scale deployment had occurred, consistent with managers making workforce decisions based on what they expect A.I. to be able to do in the near future.
In other words, junior workers are already paying the price for automation that has not yet fully arrived.
An industry-wide pattern, not just finance
The finance function is a useful lens precisely because CFOs are unusually direct about their intentions. But the pattern is not confined to finance.
A global survey of more than 850 organisations, reported by HRD Australia, found that 39% of businesses have already cut entry-level roles due to AI efficiencies in tasks including research, administration, and briefing – with a further 43% expecting to do the same within the year. In Australia specifically, 8% of employers have already stopped hiring entry-level staff entirely, and 30% have slowed their recruitment.
In the accounting and audit sector, the Big Four firms have significantly scaled back graduate hiring in recent years. In the UK, one practice leader explicitly cited generative AI as the reason for cutting approximately 200 entry-level roles, as routine audit and data-processing tasks migrate to automated systems.
Overall US programmer employment fell 27.5% between 2023 and 2025, according to Bureau of Labor Statistics data, while employment for more senior software developers fell only 0.3% in the same period.
The pattern is consistent across industries and geographies: it is the bottom of the seniority structure, not the top, that is being squeezed.
The hidden risk: Tomorrow's senior workers
The Oliver Wyman report notes, with some urgency, that 70% of CFOs plan to intensify succession planning and leadership development over the next three years. What the report does not fully reconcile is the tension between that ambition and the simultaneous erosion of the entry-level pipeline from which future senior talent has always been drawn.
Finance functions, like law firms and consulting practices, have historically operated as apprenticeship systems. Junior analysts learned by doing. They absorbed institutional knowledge, developed judgement, and moved up. As HRD has reported, contracting junior roles risks hollowing out the very pipeline that produces tomorrow's managers, specialists, and executives. Organisations may gain short-term efficiency while destroying the long-term conditions for leadership development.
The pipleline paradox
CFOs are simultaneously intensifying succession planning (70%) and reducing junior hiring (64% expect to shift away from junior roles). These two intentions are in direct tension. The senior leaders of 2035 are the junior hires of today — and today, those hires are not being made.
As HRD has reported, the perception of narrowing entry pathways is already changing what young people study and what careers they pursue. If enough of them pivot away from finance, technology, and professional services roles — the very functions now automating junior positions — employers may find themselves competing for a much shallower pool of experienced talent in five to ten years' time.
The honest question for employers
The Oliver Wyman data makes clear that CFOs are making rational decisions, function by function and firm by firm. AI does reduce the need for certain junior tasks. Headcount discipline matters. The finance pyramid does need to evolve.
But the Harvard research suggests those individually rational decisions may be aggregating into something no single employer intended: a generation of workers being locked out of the entry-level roles that have historically been the first step on the professional ladder.
The decline in junior employment appears to be driven by slower hiring rather than separations – meaning it is largely invisible in conventional unemployment statistics, which measure people who have jobs and lose them, not people who never get hired at all.
As HRD Australia has also reported, some employers that moved quickly to cut roles attributed to AI are now quietly rehiring – finding that the complexity and relationship-intensity of real work outpaced the technology's current capabilities.
The lesson is not that AI cannot replace junior tasks. Much of the evidence suggests it can, and increasingly does. The lesson is that the decision to stop hiring junior staff is not a purely technical one. It is an organisational, cultural, and ultimately strategic bet – one that will shape the talent pipeline for years after the CFO who made it has moved on.
