New data on the career outcomes of UK university graduates is offering one of the clearest early signals yet that AI is beginning to reshape demand in two fields that were, until recently, considered reliable paths to well-paid professional work: computer science and economics.
The numbers are specific enough to be striking. They are also worth understanding carefully, because what they show is not a collapse. It is something more nuanced, and in some ways more consequential.
The Coding Job Market Is Contracting at the Entry Level
The share of computer science graduates finding professional employment as coders or programmers fell from roughly 40% to just 28% in a single year, according to data compiled for the 2027 Guardian University Guide by Matt Hiely-Rayner, director of the Intelligent Metrix consultancy. The broader picture is similarly sobering: the proportion of computer science graduates moving into any graduate-level occupation dropped from over 60% two years ago to 50%.
These figures come from the Graduate Outcomes Survey, conducted by the Higher Education Statistics Agency, which gathered responses from more than 350,000 former students approximately 15 months after completing their courses in 2024.
Coding and software development were identified as the fastest-falling occupations for graduates last year. That is a notable shift in a field that has spent decades absorbing large numbers of new entrants and rewarding them with competitive salaries.
Hiely-Rayner’s interpretation is direct: with AI capable of handling what he describes as “responsive and cheap grunt work” in software development, it is difficult to avoid connecting that capability to the trend in graduate employment.
Economics Graduates Are Feeling It Too
The contraction is not limited to computer science. Economics graduates, who have historically enjoyed strong employer demand in finance, consulting, and related fields, are also seeing their prospects change. Roles in categories such as economists and management consultants showed declining demand in the data.
Charlie Ball, head of labour market intelligence for Jisc, the UK’s data and technology agency for education and research, offers a more cautious reading of causation. He notes that while something clearly happened to the software developer labour market, the evidence directly attributing it to AI remains limited. The conclusion that AI has played a role, he says, is nonetheless difficult to escape.
Ball draws a distinction worth holding onto. Many finance employers he has spoken with describe AI not as a tool for replacing jobs outright, but as something that is changing what those jobs involve. New hires are increasingly expected to arrive with familiarity with AI tools. The job exists; its shape is different.
At the same time, Ball acknowledges that employer behaviour is not uniform. Some organisations are actively embracing AI tools and reducing headcount in the expectation of gaining a competitive edge. Whether that approach persists is uncertain: as Ball puts it, everyone is experimenting right now.
Computer science graduates, for their part, appear to be adapting. There are signs, Ball notes, of diversification into adjacent roles such as cybersecurity and network engineering. Pure coding jobs seem to have been in particularly short supply in 2025, but the field has not collapsed entirely.
What This Signals About the Broader Shift
Here is what most coverage of this data misses: the story is not primarily about AI eliminating jobs. It is about AI compressing the entry point.
Graduates are, as Ball observes, the largest group of new employees in any sector. They are the ones who typically absorb the volume of routine, entry-level work that organisations need done. When AI tools become capable of handling that volume efficiently, the first group to feel the effect is not experienced professionals with years of accumulated judgment. It is the people who were about to step into those foundational roles.
This is a structural shift, not a temporary dip. The skills that made a computer science or economics degree a reliable credential for a specific type of job are being recalibrated. Employers are not necessarily hiring fewer people; they are hiring people who can work alongside AI systems, interpret their outputs, and apply judgment that the tools cannot replicate.
Universities are beginning to respond. The University of Birmingham, which rose to its highest ever position of 14th overall in the Guardian’s rankings, is offering undergraduates the option to add an intercalated year in subjects including AI and data science, without the prerequisites normally required. Adam Tickell, Birmingham’s vice-chancellor, frames this as equipping students for the world as it is, not as institutions might prefer it to be.
That framing matters. The question for students, educators, and employers alike is not whether AI will affect these fields. The data suggests it already has. The question is how quickly the educational system can adapt its outputs to match what the labour market is now asking for.
In Short
UK graduate data shows a sharp drop in computer science graduates finding coding jobs, from around 40% to 28% in a year, with economics graduates also facing reduced demand in finance and consulting roles. Experts connect the trend to AI’s growing ability to handle routine analytical and software tasks. The effect is concentrated at the entry level, where graduates absorb the most routine work. Employers are not unanimous in their response, but the direction of travel is clear: familiarity with AI is becoming a baseline expectation, not a differentiator.
Based on reporting from The Guardian - Technology.