AI impact on computer science education: schools are falling behind
The AI impact on computer science education is hard to ignore, and it’s worrying that the education system isn’t keeping up.
In the last couple of years we’ve gone from “AI can help you autocomplete” to tools that can design software, write production-ready code, debug it, and explain it back to you. Systems like Claude Code can take a plain-English description and scaffold an application in minutes.
Now compare that with what most schools are still doing in Computer Science.
At GCSE and A-level, a large part of the specification is still built around writing code by hand: algorithms, data structures, tracing, syntax, exam questions where students are assessed on producing working programs. Coding is treated as the core technical skill. That made sense when coding itself was scarce, but now we risk training people for jobs that won’t exist by the time they finish the course.
If an AI system can generate functional code on demand, the marginal value of being able to manually write a loop or implement a search algorithm is falling. Not quite to zero, because understanding logic and structure still matters. This doesn’t mean coding should disappear from the curriculum. It means coding should be framed as a way to understand systems.
I speak as a not particularly good coder, but learning machine code and programming languages helped me understand what was going on inside the hardware, and gain a far better understanding of its potential and of course limitations.
We are already seeing the impact of AI on everything from software to wealth management, so what matters now more than coding is defining problems clearly, breaking complex tasks into logical steps, and interrogating AI outputs — spotting errors, gaps, and hallucinations. That, in practice, is the AI impact on computer science education.
Perhaps the core skill we need to be teaching is knowing when not to trust the machine?
If you want to browse more of my longer-form work, start here: Field Notes. Two related pieces that touch the same “systems vs. checklists” problem from different angles are UKRI’s Applied Research Pivot: From World-Class Papers to World-Class Companies? and Bowie Bonds: IP, Cashflow, and Control for Founders.
The AI impact on computer science education
Put simply: if code production is increasingly automated, then “being able to write code” stops being the main differentiator. The differentiator becomes judgement — how you specify problems, evaluate outputs, and design reliable systems around imperfect tools. That is the AI impact on computer science education showing up in the labour market already.
What the curriculum currently rewards
The mainstream exam specs explicitly assess programming and designing/programming solutions as part of the assessment objectives and written/on-screen papers.
- GCSE Computer Science (e.g., AQA) includes assessment objectives that explicitly cover designing and programming solutions.
- A-level Computer Science (e.g., AQA) explicitly tests a student’s ability to program, including writing/adapting/extending programs in the exam.
- GCSE OCR J277 includes a full paper on “Computational thinking, algorithms and programming”, and requires algorithm-writing answers in a reference language or a high-level programming language.
What we should be teaching alongside it
Keep the foundations. But add the skills that sit above syntax:
- Problem definition (turning messy reality into a precise brief)
- Decomposition (breaking big tasks into testable chunks)
- Systems thinking (how components fail, interact, and scale)
- Evaluation (when the AI is confidently wrong)
- Data provenance and governance (what’s safe, what’s legal, what’s garbage)
- Tool orchestration (how you chain models, code, tests, docs, and deployment)
If we don’t adjust, we’ll keep producing students who can pass an exam by writing code, but struggle in the real world where the scarce skill is judgement. Again: that’s the AI impact on computer science education in plain terms.
References
AQA GCSE Computer Science (8525) scheme of assessment / assessment objectives
AQA A-level Computer Science (7517) specification at a glance (Paper 1 tests ability to program, including writing/adapting/extending programs)
OCR GCSE Computer Science (J277) specification (Paper 2: computational thinking, algorithms and programming; algorithm-writing requirements)
Ofqual GCSE subject level conditions and requirements for Computer Science
More reading (internal): Field Notes archive, Deep Tech Commercialisation.
#AI #Education #ComputerScience #FutureOfWork #EdTech #DigitalSkills
