Why teaching kids to code is key – even with generative AI

 

King Johney
3 mins read

Learning to code is essential to build cognitive skills that remain relevant as tech evolves, but can be adjusted to prepare children to leverage AI tools.

Generative AI exploded into our lives and work in 2023, with a potential compound annual growth rate (CAGR) of 42% over the next ten years and its potential impact extending to virtually all industries. It is transforming the job of computer programmers too. Already, you can build an app without writing a single line of code, and use AI to create fully-fledged video games. This all beckons an uncomfortable question: after a decade of multimillion-pound changes to school curricula, including the US government’s $4 billion investment in 2016, was teaching kids to code a waste of time?

As a website developer and programmer, I have been asked many times this year whether learning to code is still relevant. In my three years in tech exploration, I have come to understand that learning to code remains essential for nurturing cognitive skills such as problem-solving, critical thinking, logical reasoning, and creativity. These skills will remain relevant as technology evolves. In the face of concerns about AI displacing jobs and marginalising humans, coding education is more important than ever - but should be adjusted to emphasise the human skills needed to effectively leverage AI.

Claims that kids no longer need to learn to code - like those made by the OECD’s education chief Andreas Schleicher - misunderstand the nature of software engineering. Non-techies may assume that a coder’s job is mainly to code. Even prior to the rise of AI, the job of a software engineer focused less on actually writing code than on using libraries of pre-written code, analysing what the code does and adjusting it for purpose. Increasingly, that pre-written code is now generated by an AI. In both cases, the engineer’s role is mainly about critical thinking and problem-solving.

How do young people build critical thinking and problem-solving skills? Steve Jobs' decade-old insight that “everyone should learn to code because it teaches you how to think” remains as valid as ever. A 2019 study found that one month of computational thinking and coding activities is equal to seven months of standard maths and science for developing executive functioning in children's brains.

One teaching method that prepares children to leverage AI, called PRIMM (Predict, Run, Investigate, Modify, Make), involves getting students to review pre-written code, predict its behaviour, run and analyse it, make adjustments, and finally, create something new based on what they've learned. Twelve-year-olds might investigate block-based code for an animated dance party with a ghost, anticipating the ghost's movements based on the code, fixing bugs, and then adding new elements. Such comprehension and critical analysis skills build foundations for working with AI-generated code.

Learning to code is still the most efficient way to build the cognitive skills needed to effectively leverage AI tools and take advantage of emerging job opportunities like “AI prompt engineering”, identified by the World Economic Forum as one of top 3 jobs of 2024. Now where is that more important than in underserved communities, as the gap widens between those who can harness AI for advancement and those lagging desperately far behind.