When Noah Ross Grundmann finishes his accounting degree and steps into full-time work, he knows his future bosses won’t be looking for an AI newbie.

“Students need to graduate understanding how AI applies to real workplaces and where compliance, data privacy, accuracy and human accountability actually matter,” said the 20-year-old, who also works as a finance and compliance officer at a not-for-profit.

AI expertise has become increasingly crucial for university leavers’ job prospects yet it is a double-edged sword. The technology is so capable that it could wipe out half of all white-collar entry-level jobs, says Dario Amodei, the chief executive of Anthropic, which built AI assistant Claude, while the World Economic Forum predicts that 39 per cent of jobs will “fundamentally change” by 2030.

Those predictions are being realised already. Law firm MinterEllison in May announced it was cutting its graduate intake from more than 100 to 72, The Australian Financial Review reported.

“Responsible use of AI is improving efficiency in some of the more routine work graduates traditionally start on, while demand continues to shift towards complex matters,” MinterEllison chief people officer Rachel Banks said. “That combination has led us to take a more targeted approach.”

Other law firms said their reduced graduate hire was not because of AI but a reflection of the wider labour market.

Changes are also afoot in accounting. At PricewaterhouseCoopers, AI has changed the Australian graduate hiring pattern. The firm has cut entry-level roles by 6 per cent, while it is looking for half of its 2027 graduate intake to come from non-accounting backgrounds. Deloitte hired 655 university leavers this year (down from 730 last year), with “strong demand” for technology, data and engineering graduates.

Morgan Stanley’s US-based AI disruption tracker found that while the technology has not harmed the job market too much, younger workers bear the greatest impact.

“Entry-level roles, which have historically served as the training ground for graduates, are increasingly being reshaped or removed altogether as AI takes on routine cognitive tasks,” said business trend forecaster Michael McQueen. “It’s almost as if the first few rungs of the career ladder are being removed in a number of white-collar professions.”

He said the short-term outlook was “more competitive and uncertain” due to the cooling labour market, employers looking to AI for tasks previously done by junior workers, and growing scepticism about the value of a degree as education costs rise.

“In the next 12 to 24 months we are likely to see a tougher transition from study to work, particularly for those relying on traditional pathways.”

Students and employers alike want universities to move away from policing AI and towards teaching AI skills for the workplace.

“AI literacy is rapidly becoming a core workforce capability. This means teaching students how to critically evaluate outputs and apply them strategically and ethically,” RMIT Online director of customer success Anshu Arora said.

“Young workers are particularly vulnerable. Gen Z are twice as likely to overstate their AI skills compared to Baby Boomers, which increases the risk of misuse if they are not equipped with strong critical evaluation skills.”

The Future Skills Organisation said employers report graduates are starting work lacking practical skills, AI know-how, communication ability and professional judgment.

“As a result, employers are bearing the cost of developing workforce readiness post-hire. This takes away employer investment from early career development for graduates,” it said in a submission to the Australian University Graduates Senate inquiry.

Beyond AI skills, McQueen said graduates could stand out in a difficult market if they “double down on the human skills that are harder to automate, such as communication, curiosity, creativity, ethical reasoning and collaboration”.