The RBA’s monitoring
According to a report by IT之家, citing THE DECODER, AI has not yet significantly shaken the entire Australian labor market, but the situation is different for those entering the workforce. Mike Major, a researcher at the Reserve Bank of Australia (RBA), has developed a monitoring system for the occupations most exposed to AI, including programming, office tasks and customer support.
Since the release of ChatGPT at the end of 2022, job postings have fallen in every category, with the steepest reductions precisely in the occupations with the greatest theoretical exposure to AI. In this group, workers aged 15 to 24 and recent graduates show lower employment and higher unemployment than their older colleagues.
Major himself, however, urges caution: the research does not demonstrate a direct causal relationship between AI and the deterioration in employment conditions, and it measures the theoretical exposure of occupations, not companies’ actual use of AI. The post-pandemic economic cycle and changes in remote work may also be weighing on the data.
Partial confirmations and the Anthropic case
The Australian picture finds echoes elsewhere. In 2025, a Stanford study based on millions of U.S. payroll records found a relative 13 percent reduction in employment among 22- to 25-year-olds in highly exposed occupations, with a decline of close to 20 percent for young software developers. In the same sector, more experienced workers remained stable or saw their position improve.
Anthropic, cited in the material, observed a decline in opportunities for 22-25-year-olds in exposed occupations, equivalent to about 14 percent lower likelihood of finding work, but without a systematic increase in overall unemployment. The company also disclosed that, since the beginning of 2025, the share of junior new hires in technical roles has fallen from 27 to 9 percent, and that by August 2026 its AI agents had accumulated an execution time equal to six times the hours worked by human employees.
Anthropic itself acknowledges that its own case, as a software company heavily dependent on AI, is not representative of the entire market. Researchers speak of the risk of a “tragedy of the cognitive commons”: by reducing junior hiring, the pipeline that trains the expert professionals of the future could be broken—assuming such professionals are still needed in the future.
Aggregate effects remain contained
In Denmark, a study of 25,000 workers found 64 percent use of AI with no measurable changes in wages or hours, with an average weekly saving equal to 2.8 percent of working time. However, a January 2026 analysis of 10.6 million LinkedIn profiles and U.S. unemployment data, led by Morgan Frank of the University of Pittsburgh, placed the onset of employment difficulties in exposed occupations at the beginning of 2022, before ChatGPT spread, indicating a greater role for macroeconomic factors.
At the same time, opposite signals are emerging. According to the World Economic Forum’s Future of Jobs 2025 report, cited by the source, 77 percent of surveyed companies prioritize reskilling over cuts, and 78 million net additional jobs are expected by 2030; on LinkedIn, AI-related job postings have increased sixfold in a year. It therefore remains to be seen whether the decline in entry-level positions is a lasting effect of automation or the result of cyclical factors: the available material does not make it possible to determine this.
