19 June 2026
Every few years, education is visited by a familiar prediction: an aspect of a subject is declared obsolete, overtaken by technological advance. Recently, that spotlight has fallen on coding. With AI tools now able to generate functional code in seconds, some are asking a seemingly logical question: if machines can code for us, why should students learn to do it themselves?
It’s an appealing argument, particularly in a profession that has weathered repeated waves of technological enthusiasm, but it is also fundamentally flawed. Not because AI won’t reshape coding — it undoubtedly will — but because coding has never been solely about producing future programmers.
If anything, AI strengthens the case for teaching coding, rather than weakening it.
Coding as a way of thinking
One of the most persistent misunderstandings about coding is that it is simply a technical skill — a matter of writing commands and producing working programs. In reality, coding is better understood as a form of cognitive training.
When students learn to code, they are developing a set of mental habits that extend far beyond the screen: problem solving, logical reasoning, hypothesis testing, creativity within constraints, and resilience through iteration. A bug is not just a mistake; it becomes a prompt for investigation. Students are required to test ideas, refine their thinking, and persist through failure.
Coding functions almost as a “prosthesis” for thinking — extending a student’s ability to engage with difficult ideas. It creates an environment where trial and error is not only accepted but encouraged, helping students to build confidence and capability.
These are not incidental by-products of coding. They are the point.
Importantly, these are precisely the skills that will matter most in a world shaped by AI. As tools become more capable of producing outputs instantly, the human role shifts towards evaluating, questioning, and improving those outputs. Coding cultivates the kind of thinking required to do exactly that.
In an AI-infused world, the question is not whether students can generate answers — but whether they can understand and assess them.
Curriculum reform isn’t retreating — it’s reinforcing
Recent curriculum developments underline this point. The 2025 Curriculum and Assessment Review makes it clear that computing — including programming — is not being phased out. Instead, it is being strengthened.
The rationale is straightforward: digital skills are essential for participation in modern society. From healthcare to law, logistics to the creative industries, technology underpins every sector of the economy. Students need to understand not just how to use technology, but how it works.
The planned 2028 curriculum reflects this shift. A broader, more future-facing Computing GCSE will replace the narrower Computer Science qualification, embedding programming within a wider framework of digital literacy, data, systems thinking, and computational reasoning. At the same time, proposals for new qualifications in AI and data science further reinforce the importance of coding as a foundational skill.
This is not a retreat from coding. It is a recontextualisation — recognising that coding is part of a larger ecosystem of digital understanding.
The workforce problem no one is talking about
Beneath debates about AI replacing programmers lies a quieter, longer-term risk: the erosion of the talent pipeline.
Every experienced developer begins as a novice. Those early stages — writing basic programs, making mistakes, learning through practice — are essential. If AI begins to replace entry-level coding tasks, and education responds by scaling back programming, we risk creating a generation with limited understanding of how systems actually work.
AI can generate code, but it cannot take responsibility for it. It cannot fully ensure that systems are secure, ethical, or sustainable. It cannot fully anticipate the societal consequences of the technologies it helps create.
Human expertise will still be required — but that expertise must be developed over time. Without continued emphasis on coding education, that pipeline may falter.
More than code: the meta-skills that matter most
When students write programs, they are not just learning syntax. They are developing transferable skills that will shape their broader lives: attention to detail, abstraction, decomposition, adaptability, and evaluation.
These meta-skills are increasingly valuable in a world where information is abundant and rapidly generated. The ability to approach unfamiliar problems methodically, to break them down, and to refine solutions is vital — not just in computing, but in everyday decision-making.
Evidence suggests that students who engage with coding from an early age often demonstrate stronger analytical and problem-solving abilities. More importantly, they develop ways of thinking that persist into adulthood.
In a landscape where AI can produce answers at speed, human value lies in asking the right questions — and in recognising what makes a good answer.
So, should we still teach coding?
The real question is not whether AI can write code. It increasingly can. The more important question is whether students will understand the systems that shape their lives, think critically about the tools they use, and develop the cognitive flexibility needed to navigate change.
Coding remains central not because every student will become a programmer, but because no student can afford to be digitally illiterate.
AI may transform how software is written. It will not replace the human capacities that coding develops: thinking, questioning, creating, adapting. Those remain at the heart of progress — and will continue to be so.
Rethinking assessment: the future of the NEA
If coding is to remain central, then assessment must evolve alongside it. Nowhere is this more evident than in the future of the A level Computer Science non-examined assessment (NEA).
Rather than replacing the NEA with a written exam — and losing the creativity and practical programming it fosters — there is a compelling case for reimagining it to reflect the realities of modern development. AI should not be excluded; it should be embraced as part of the process.
A future NEA could place AI at the heart of problem-solving, with students demonstrating how they work alongside these tools rather than independently of them. Instead of producing lengthy written reports documenting every stage, students might maintain a concise, structured development diary — potentially even as a video record — capturing their thinking and decision-making.
This process could be organised around a series of stages we call. “CODEAI”:
- Create: Developing an initial solution to a problem.
- Orchestrate: Collaborating with others and with AI tools.
- Debug: Integrating components and resolving issues.
- Experiment: Exploring alternative approaches and improvements.
- Adapt: Refining and modifying generated code.
- Improve: Enhing the final solution to better meet user needs.
Such an approach preserves what makes the NEA valuable — creativity, independence, and problem-solving — while aligning it with the evolving reality of coding in an AI-supported world.
Coding is not disappearing. It is changing — and education must change with it.
By keeping coding at the heart of the curriculum and adapting how we teach and assess it, we ensure that students are not just passive consumers of technology, but active, critical participants in shaping its future.
