“Machines are taking our jobs” is a familiar fear, often met with a reassuring claim that while some jobs vanish, new ones will emerge to replace them. Today, that reassurance is harder to sustain.
Annie Phan
ESSA Monash Clayton
[Annie Phan is a second-year Law and Commerce student majoring in Economics, with a strong interest in global economic systems and international relations. She hopes to deepen her understanding of how economic incentives shape human behaviour and drive real-world outcomes across societies.]
Disclaimer: The views expressed in this article are those of the author and do not necessarily reflect the views of affiliated organisations.
AI is reshaping entry‑level work, with algorithms able to perform junior tasks in seconds. The structural question is simple: how does the next generation gain experience when the traditional “foot in the door” roles are disappearing?
Ask two economists that question and you’ll end up with two equally valid yet completely different answers. Not because either is mistaken, but because each is evaluating a different aspect of the problem.
Companies that have embraced AI investments show a positive pattern. A study of 21,559 U.S. firms by Ramp’s Economics Lab and Revelio Labs used actual AI spending data rather than survey responses and linked it to verified workforce records. The finding reveals that the firms investing the most in AI increased their entry‑level hiring by 12 percent over the next two years.1
When we look across the broader labour market and focus on young workers trying to enter the most AI‑exposed occupations, the pattern becomes more concerning. Anthropic’s economists find that job‑finding rates for 22‑ to 25‑year‑olds fell by about 14 percent after ChatGPT’s launch, with no comparable decline among workers over 25.2
So which study should we believe and be concerned about? Both. Each one answers a different question. One examines what happens inside firms that invest heavily in AI. The other looks at what happens across the wider economy for young people trying to enter AI‑exposed jobs. That difference in focus explains much of what follows, and it shows why technological unemployment rarely has a single, simple answer. Three ideas from labour economics can help make sense of this.
Idea 1: AI can replace a task, or it can help with one
Think of any job as a bundle of tasks that AI can either take over some or help a worker complete them more quickly. Economists describe this difference as substitution versus complementarity: the AI either replaces the worker for a task or supports them in doing it. Claude operates similarly, with about half of its conversations showing automation,3 where the AI completes the task itself, and the other half showing augmentation, where the AI works with the user to refine the task.4
Substitution and complementarity leads to opposite hiring effects. When AI substitutes for a task, it competes with the entry-level worker, so hiring falls. When AI complements a task, it boosts the worker’s capability rather than making them redundant, so hiring should remain stable or increase.5
Idea 2: young workers feel it first, even when the economy doesn’t
Labour economics explains that firing is costly, but firms can stop hiring young workers at no cost, so the impact of AI on employment shows up first in entry‑level hiring.6 Severance, sunk training costs, morale concerns, and the loss of firm‑specific knowledge make companies reluctant to fire even when tasks are automated. By contrast, choosing not to open a junior role is frictionless, so firms reduce entry‑level labour through quiet hiring freezes rather than layoffs.7
Idea 3: entry‑level jobs were always partly about training
No one hires a junior employee for what they can do today, rather they are hired for what they’ll be able to do in a year. Therefore workers are treated as an investment, where the trade-off is to accept lower productivity now in exchange for more later.8
If AI can produce that early-career output directly, this investment stops making sense. Why invest in a junior worker’s early-stage learning curve when AI can deliver the output immediately? We can’t directly confirm firms think this way, but the pattern in their behaviour is consistent with it. Across more than a billion job postings worldwide, AI‑exposed junior roles are now seven times more likely to require senior‑level skills such as judgment and leadership.9 This reflects skill-biased technical change, where new technology raises the value to some skills and reduces it for others.
What this looks like in Australia
Australia’s data remains inconclusive, but corporate hiring trends are starting to tell a different story. According to Indeed’s Hiring Lab report, graduate opportunities fell by 15 percent in 2025.10 Occupations most exposed to AI dropped more sharply, and entry-level roles tend to be more vulnerable to economic cycles.11
The figures behind these hiring trends are exemplified when, for example, Atlassian reduced its headcount by 1,600 worldwide, of which 500 were Australians.12 Kris Grant, a recruitment executive at ASPL Group, attributes these trends to lower-value tasks being automated while other roles are being redesigned.13 Because automation follows the same substitution mechanism that removes the lowest‑value tasks first, the impact concentrates at the entry level where routine work is replaced fastest.
When entry-level stops being entry-level
It’s not a crisis, but it is a clear shift. Entry‑level roles in AI‑exposed fields are becoming more selective and require a different skill mix. Some of that change reflects AI, and some reflects a broader hiring slowdown hitting the same cohort, and the data can only partly separate the two.
A degree no longer opens the door on its own. Employers, AI, and new graduates are collectively rewriting what does.
References
[1] Simon, Lisa, and Ryan Stevens. 2026. “A New Look at AI’s Impact on Jobs: Firm-Level AI Spending and Workforce Adjustment.” Ramp Economics Lab and Revelio Labs, June 30, 2026. https://ramp.com/data/ai-jobs-impact. [2] Massenkoff, Maxim, and Peter McCrory. 2026. “Labor Market Impacts of AI: A New Measure and Early Evidence.” Anthropic, March 5, 2026. https://www.anthropic.com/research/labor-market-impacts.[3] Anthropic. 2026. “Anthropic Economic Index Report: Economic Primitives.” Anthropic, January 15, 2026. https://www.anthropic.com/research/anthropic-economic-index-january-2026-report.
[4] Anthropic. 2025. ‘Introducing the Anthropic Economic Index’. Anthropic.Com. https://www.anthropic.com/news/the-anthropic-economic-index.
[5] Burnham, Kristin. 2026. ‘Pro-Worker AI, Explained’. MIT Sloan, June 17. https://mitsloan.mit.edu/ideas-made-to-matter/pro-worker-ai-explained.
[6] Brynjolfsson, Erik. 2025. ‘Canaries in the Coal Mine? Six Facts About the Recent Employment Effects of Artificial Intelligence – Stanford Digital Economy Lab’. Stanford Digital Economy Lab, December 23. https://digitaleconomy.stanford.edu/publication/canaries-in-the-coal-mine-six-facts-about-the-recent-employment-effects-of-artificial-intelligence/.
[7] J. Sucher, Sandra. 2024. ‘The Hidden Costs of Layoffs’. Harvard.Edu, November 20. https://www.hks.harvard.edu/centers/mrcbg/programs/growthpolicy/hidden-costs-layoffs.
[8] Hewes, Caitlin. 2026. ‘If You Replace Junior Talent With AI Today, Who Will Lead Your Company Tomorrow?’. Forbes, July 29. https://www.forbes.com/councils/forbeshumanresourcescouncil/2026/07/29/if-you-replace-junior-talent-with-ai-today-who-will-lead-your-company-tomorrow/.
[9] PricewaterhouseCoopers. 2026. ‘AI Reshapes Global Labour Market Into Two Distinct Paths, Rewarding Human Skills: PwC 2026 Global AI Jobs Barometer’. PwC. https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-jobs-barometer.html.
[10] Indeed Hiring Lab Australia. 2026. “Nice Try, AI: Australian Graduates Are Still Getting Hired.” Indeed Hiring Lab, April 22, 2026. https://www.hiringlab.org/au/blog/2026/04/22/nice-try-ai-australian-graduates-are-still-getting-hired/. [11] Australian HR Institute. 2026. “Graduate Hiring Remains Stable Despite AI Disruption, Research Finds.” AHRI, May 5, 2026. https://www.ahri.com.au/articles/graduate-hiring-remains-stable-despite-ai-disruption-research-finds. [12] Knight, Elizabeth. 2026. “The ‘Race to the Bottom’ on AI Job Cuts Has Already Started.” Sydney Morning Herald, April 9, 2026. https://www.smh.com.au/business/companies/the-race-to-the-bottom-on-ai-job-cuts-has-already-started-20260409-p5zmi5.html. [13] Hendy, Nina. 2026. ‘AI Is Taking Entry-Level Jobs. For Young Workers, It’s a Huge Problem’. The Sydney Morning Herald, April 16. https://www.smh.com.au/business/workplace/ai-is-taking-entry-level-jobs-for-young-workers-it-s-a-huge-problem-20260416-p5zohb.html.
