Graduate jobs are supposed to offer two things: a salary and a way to learn. Much of that learning has traditionally happened through repetitive work. Junior employees research a company, prepare a first draft, review documents or analyse figures, then receive feedback from someone more experienced.
Generative AI is increasingly capable of assisting with those same tasks. That creates a difficult question for employers and graduates alike: if the first draft takes minutes, how does someone learn to judge it?
Is AI already reducing entry-level hiring?
There are reasons for concern, but the evidence is not a settled verdict. Anthropic's March 2026 labour-market research found no systematic increase in unemployment in highly exposed occupations within its analysis. It did, however, identify suggestive evidence of slower hiring for younger workers in exposed occupations.
That is not proof that AI caused a particular recruitment freeze. Economic conditions, hiring cycles, sector weakness and changing business models can also affect junior vacancies.
The honest position is that entry-level work may change significantly, and some early signals merit attention, but exact forecasts of disappearing graduate jobs would be premature.
Which graduate tasks are changing?
In professional services, AI can assist with desk research, document summaries, slide outlines, spreadsheet explanations, first drafts of communications and basic code.
These are common starter tasks across banking, accounting, law, consulting and technology companies. The risk is not just that employers need fewer hours to complete them. It is that junior staff may receive less deliberate training in the foundations.
A graduate who accepts an AI-generated answer without understanding the source, calculations or assumptions has not acquired the expertise the task was supposed to teach.
What employers should do differently
The sensible response is not to give new joiners a chatbot and call it training. Junior employees need access to the underlying work, clear review standards, meaningful feedback and opportunities to practise without shortcuts.
Managers might ask graduates to complete a small analysis manually before comparing it with an AI-assisted version. They could review not only the finished answer but also the evidence and judgement behind it.
The goal is to produce professionals who can recognise a poor answer, whether it was written by a person or a machine.
How to make yourself more employable as a graduate
Show an example of work, not a list of tools
Choose a public business question relevant to your target role. Build a short, source-backed analysis, check it thoroughly and explain your conclusion in plain English.
Learn to challenge AI output
Ask an approved tool for an initial answer, then identify unsupported claims, missing information and alternative interpretations. In interviews, explaining your verification process can be more convincing than claiming you are an "AI expert".
Strengthen the fundamentals
An aspiring accountant should understand financial statements. A future lawyer should know how to research authority. A graduate consultant should be able to structure a problem. AI is more useful when you can tell whether its work is good.
Develop the human side early
Ask sharper questions, communicate clearly, take feedback and build relationships. Those skills are easier to demonstrate when paired with strong technical work.
What to ask at an interview
Ask how the team uses AI, which tools are approved, how junior work is reviewed and how employees build expertise if more preparation is automated. These questions show commercial awareness without sounding frightened of technology.
If you are already employed, use the 30-day plan for AI at work. If you are still choosing a path, read our guide to the white-collar tasks most exposed to AI.
