Are UK universities preparing software engineers for an AI-transformed profession? PILOT
AI is transforming how software is built, tested, delivered, and maintained. Universities are responding with new AI degrees and expanded machine learning modules. But almost nobody is asking the harder question: are the core programmes that produce the UK’s working software engineers adapting to a profession that AI is reshaping from the ground up?
This site presents a structured, evidence-based assessment of 23 UK university computer science and software engineering programmes. It asks not whether they teach AI as a subject — most do — but whether they are preparing graduates for the practical reality of AI-augmented software development.
The answer, based on publicly available evidence, is sobering. Zero institutions in our sample require students to use AI coding assistants as part of assessed software engineering work. The profession has moved; the curriculum has not followed.
Three things you should know
Computer science is not software engineering
When a CS department says it is “all over AI,” it typically means strong research in machine learning and new AI degree programmes. That is valuable, but it is not the same as rethinking how software engineering is taught. The AI transformation of software development is primarily a software engineering question — how teams work, how quality is assured, how systems are architected — and it is falling between institutional cracks.
There is a two-speed system
Oxford has ChatGPT Edu, Copilot, and Gemini for all staff and students. At none of the four teaching-focused institutions in our sample could we find evidence of institutional AI tool provision. These post-92 universities produce a significant share of the UK’s working software engineers. Their graduates face the same AI-transformed workplaces as Oxford’s — but with different preparation.
The strongest findings come from unexpected places
UCL’s “Automated Software Engineering” module is the only one in the sample that explicitly teaches using LLMs for SE tasks. Ulster’s MSc has the most explicit MLOps content — more specific than anything at Oxford, Cambridge, or Imperial. Cardiff’s National Software Academy maps to SFIA professional skills. Queen’s Belfast has mandatory paid year-long placements with 500+ employers. These are not the institutions that dominate the headlines.
Explore this assessment
What the data reveals
Detailed commentary on findings across 23 institutions, organised by the patterns that matter.
Read the analysis →Interactive data
Explore the heatmap, compare institutions by dimension, and drill into individual profiles.
Open the dashboard →How this was built
The five-dimension framework, data sources, rating scale, and why we reframed the question.
See the methodology →The sample at a glance
23 institutions across three tiers and all four UK nations, each assessed against five dimensions of readiness for AI-transformed software engineering practice.