Choosing a university degree has never been straightforward. Students and parents have always asked which courses and majors offer the best returns and lead to secure careers.

But now the age of generative artificial intelligence, bringing powerful new tools such as ChatGPT and Claude, has made those decisions much more complex.

In New Zealand and elsewhere, students are finding it harder than ever to judge which careers will thrive, which skills will remain valuable and how AI might disrupt the professions they hope to enter.

Generative AI can already draft reports, summarise research, write code, analyse data and generate professional-looking content. Although not yet reliable enough to replace professional expertise, it is changing the kinds of work graduates are expected to do, particularly at the start of their careers.

A recent survey of New Zealand business leaders found 87% of organisations had seen job roles change or disappear because of AI, while one third reported slowing entry-level hiring.

This all creates an awkward question for universities: if AI can perform many of the tasks once used to demonstrate graduate competence, what should a degree now provide?

We argue the answer is not to abandon disciplines, nor for every student to become a computer scientist.

Instead, universities should reconsider how early students specialise – and whether current programmes provide genuine opportunities for cross-disciplinary learning.

Why the boundaries are blurring

The biggest challenges facing society rarely fit neatly within a single discipline.

Climate change involves economics, law, politics, communication and justice, while public health draws on medicine, logistics, behaviour, ethics, data and public trust.

Generative AI makes it easier to draw on knowledge from neighbouring disciplines. But it also makes superficial understanding easier to disguise.

A polished answer can conceal weak reasoning, missing evidence or invented facts. That makes the ability to question, verify, interpret and take responsibility for information more valuable than simply producing it.

Disciplines remain important because they teach more than content. They provide ways of testing evidence, evaluating claims and deciding whether a conclusion is sound.

An accountant must know whether an analysis is defensible. A lawyer must understand authority and precedent. An engineer must judge whether a design is safe.

Generative AI can produce work that resembles that of these professions. It still struggles to determine whether the result is accurate, responsible or fit for purpose.

As AI makes it easier to produce convincing answers, the ability to question assumptions, evaluate evidence and exercise judgement becomes more important.

The challenge for universities is therefore not to choose between specialisation and generalism, but to combine the strengths of both.

Graduates with broad but shallow knowledge may struggle in high-stakes settings. But those with expertise confined to a single discipline may also find it harder to navigate increasingly interconnected professions.

The capabilities employers value – critical thinking, professional judgement, communication, teamwork, ethical reasoning, adaptability and problem-solving – cut across disciplines. They are developed in different ways, but applied across countless occupations.

A consulting firm may recruit graduates from commerce, computer science, psychology or journalism. Their expertise differs, but all must analyse unfamiliar problems, work across disciplines and exercise sound judgement.

Business schools illustrates this. Issues such as AI adoption, sustainability and pricing apply to areas ranging from finance, technology and consumer behaviour to ethics, regulation and organisational change.

Depth, breadth and adaptability

So, what path should universities take when redesigning courses? One useful model is the “T-shaped” graduate.

The vertical stroke represents depth: real expertise in a discipline or profession. The horizontal stroke represents breadth: the ability to communicate across fields, use generative AI critically and make decisions under uncertainty.

Students could begin with a broader foundation that integrates disciplinary knowledge with AI literacy, ethical reasoning, communication and collaborative problem-solving.

Specialisation may come later, once students understand how fields connect and why complex problems require multiple perspectives.

Universities could also create shared learning across faculties. Business, design and computer science students might work together on responsible AI adoption.

Law, health and data students might examine privacy and automated decision-making. These experiences would help students understand what their discipline contributes, where its limits lie and when other expertise is needed.

Universities should treat AI as a core form of literacy, rather than something covered in a short workshop on writing prompts.

Students need to understand what these systems can and cannot do, how errors and bias arise, how privacy can be compromised, and how AI-assisted work should be evaluated.

Assessment should place less emphasis on producing answers and more on explaining reasoning, defending decisions and responding to new evidence.

As information becomes easier to generate, the value of a degree will increasingly lie in the judgement to use it well.

There may be no such thing as an AI-proof degree. The degrees that endure will equip graduates with deep disciplinary expertise, the breadth to work across fields and the adaptability to keep learning as technology evolves.

The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.

Read More: Read More