If AI writes code, is coding still crucial for kids to learn?
For years, coding has been thought of as a useful skill for children to learn. It’s integrated into computer science classes and a number of organizations are dedicated to helping kids code.
But now, AI tools can write code themselves.
Marketplace’s Stephanie Hughes spoke with Monica McGill of the Institute for Advancing Computing Education about what the expanding capabilities of artificial intelligence mean for coding as a necessary — or not so necessary — skill.
The following is an edited transcript of their conversation.
Monica McGill: There was a report in 2021 that came out from McKinsey & Co., basically stating all future workers need programming literacy skills and computational thinking skills, and they’re both underlying AI. You can’t really understand the mechanisms under AI if you don’t understand computer science, computational thinking, programming and even data science. So yes, this is before ChatGPT really launched and everyone jumped into that. However, it’s still not going to take away [the fact] that if you don’t understand how it’s working, if you don’t understand what it’s producing, then you can’t really say what it’s doing is what you want it to do.
Stephanie Hughes: Is the use of AI in coding having any effect on coding education right now?
McGill: Yes. We know that students are using AI, and teachers are a little bit more reluctant, but they’re starting to adopt it as well. So it’s important to actually have students use AI in the classroom as well when they’re learning programming, so that they can see what’s happening and the teacher can explain, “OK, this is correct, this is incorrect. Let’s investigate why it is or isn’t correct.” That can help train students so that when they are out, suppose they do have a career that’s actually in programming, they might be able to approach AI a little bit differently and interrogate in a way that maybe newer folks who have not had that type of training are able to do.
Hughes: Is there a way for young programmers to use AI as a tool, but also, you know, understand the fundamentals of what they’re doing and learn to judge the AI’s code?
McGill: Yeah, I think that that is one of the areas that at least our research institute is looking at, really trying to understand, what does it mean to not only teach young programmers about AI and using it in programming, but also having it assist with development and the learning? But we don’t know yet how that is all going to play out. There’s a practice in teaching young students programming called pair programming. So you pair up two learners, two novice learners, and they work together to solve problems that, for example, are presented to them in the classroom. So we know that that can work for some students better than others, but what happens when you take one of those students and you pair them with an AI system that replaces the other pair programmer, so to speak? So how does that training change? How does the learning change for the student? Is it better, is it worse? We don’t know yet, but we’re going to keep investing and getting that as we go along.
Hughes: You know, there’s a lot of things we teach in schools that we don’t use later. Like, not everyone who takes computer science will grow up to become a computer scientist, but we teach these things because they develop our skills. Do you have a sense of whether coding could become less of a practical skill but more of a foundational one in the future?
McGill: Yeah, I think it should be already. If you look at the countries with the highest [gross domestic product], two of the countries that stand out are Japan and the U.K., in particular England. So both Japan and England have mandatory K-12 computer science education requirements for all students. And the U.S. does not, right? And across all 50 states, it varies of what requirements are for students, and that’s an area where we could really focus on, if you can get it into these earlier grades. And that will then support what the McKinsey report came out and said, that all future workers, no matter what their role is, need to have some understanding about programming and computational thinking and data science because it basically touches everybody’s work. You can’t have a job that doesn’t in some way use computer science.
The New York Times has a piece exploring how some software developers are turning to AI more frequently to help them code. It quotes the leader of the developer site GitHub, who says the role of the human developer will evolve to become “the conductor of an AI-empowered orchestra.”
One impetus for this story was a blog post by a developer, Namanyay Goel, lamenting that many junior programmers are relying on AI chatbots to write code.
It enables them to work superfast. But when he asks them why the code works the way it does, Goel says he just gets blank stares. He calls it “trading deep understanding for quick fixes.”
He suggests developers interrogate the AI they’re working with, ask it how it got the answer it did. He writes, “Sure, it takes longer, but that’s literally the point.”
To put it another way, in the immortal words of singer Miley Cyrus: “It’s the climb.”
Stories You Might Like
The future of this podcast starts with you.
Every day, the “Marketplace Tech” team demystifies the digital economy with stories that explore more than just Big Tech. We’re committed to covering topics that matter to you and the world around us, diving deep into how technology intersects with climate change, inequity, and disinformation.
As part of a nonprofit newsroom, we’re counting on listeners like you to keep this public service paywall-free and available to all.
Support “Marketplace Tech” in any amount today and become a partner in our mission.
