The focus on teaching students how to code—a big emphasis for years—is now expanding to showing them how artificial intelligence works.
Code.org, one of the major K-12 computer science education curriculum providers, is rebranding to CodeAI, expanding its mission from computer science education into learning about AI and building “digital fluency,” the nonprofit announced this month.
Since the public release of generative artificial intelligence tool ChatGPT, the Seattle-based nonprofit has been providing educators with resources more focused on teaching about AI than coding. In December, it rebranded its Hour of Code, an annual event designed to get students interested in coding and computer programming skills, to Hour of AI, which teaches kids how that technology works.
The rebranding comes amid concerns about how AI is threatening careers in computer science and questions around whether computer science education needs to change. One big concern among many students who are interested in computer science careers and people already working in the field is that AI can write code on its own.
There have also been growing concerns over AI’s effects on students’ critical thinking skills and mental health, and the amount of time kids are spending on screens in the classroom.
Education Week spoke with Karim Meghji, the CEO of CodeAI, about the rebrand, concerns around AI, and the future of computer science education.
The interview has been edited for length and clarity.
Why is Code.org rebranding to CodeAI?
Shifting from this computer science domain to a broader surface area of digital sciences, where AI science and data science increasingly are important in rounding out these areas of digital technology that every student should have access to, is the impetus for this.
We’ve learned over the last three years the importance of teaching students about AI, because we’ve been teaching AI for multiple years. The rebrand was less about a forecast of things that we have yet to learn and do. It was more acknowledgement of the work we’ve done and the learnings we’ve had about the importance of not just teaching students how to use AI tools, but how do they actually work, and the fact that when you teach students about how they work, they’re not only better consumers in the digital world around them, but it starts to give them the agency to actually control it, direct it, affect it, build with it.
These are the things that we want for these students in whatever pursuit they have, whether they choose to pursue a physical sciences career and build things in the real world, or build things in digital engineering—all of these different domains can benefit from that.
Is your organization’s mission changing?
Our mission hasn’t changed. Technical education is what we’re focused on, and with the idea of giving every student that opportunity, but the mix of what’s in there is what’s changing. The definition is expanding, and we’re rebalancing what’s in technical/digital education.
What can educators expect from CodeAI?
Increasingly, our curriculum will include AI science. How do models work under the hood? What is actually happening when a student prompts before they get a response? All the way down to, potentially, even the hardware layer. Why are GPUs [graphics processing units, a specialized electronic circuit] used? What are the probabilities in mathematics that go on under the hood? Demystify the machine to the level [where] students, after they understand how these things work, see [the technology] differently. They understand a little bit more what’s happening. They don’t see [it] as magic.
We want to get students hands-on with models. Part of learning how this technology works is not just one-way communication. It’s not videos and lectures. It is touching the technology, trying things, training a model, and seeing how different inputs affect the outputs.
We’ll continue to invest in computer science education. Computational thinking as a foundational, durable skill continues to be important. Learning about how code works so that you can evaluate AI outputs will continue to be important.
Then we’ll round that out with more data science, because I do think in a world with AI models consuming data as fuel, we’ve got to teach students about data, how to construct data, where data is sourced from, how to prepare it for models, how it affects models.
How is CodeAI addressing concerns around AI’s impact on society and human interaction?
The question about responsible computing has been around for decades. Data centers have been sitting around at scale, generating heat and occupying space for decades. But we need to continue that conversation. We need to be cognizant of how computing systems affect society, both from an ethical and a social perspective.
The other area that is thematic for us is this idea of how humans and computers interact. It’s something we’ve taught about for some time. With more students becoming builders with AI, we see that as a more important skill to teach about.
Our educational materials and our approach has not been just to teach facts and knowledge and skills. We teach a lot of mindset. Mindsets are very important in this world. One thing you’ll see us do even more is present things into the classroom where teachers and students can actually debate, critically think, about a thing, whether it’s the social impact or how human-computer interactions will evolve.
There’s a lot of discussion about how AI is going to erode critical thinking. I actually think it’s the opposite. With good pedagogy, you can enhance critical thinking in a world of AI, not simply go to cognitive offloading. I think partially that’s because we haven’t gotten yet to curriculum rethinking and pedagogical improvement in a world of AI. We’re thinking a lot about that, and I hope our colleagues in the other domains will do the same.
Can you talk more about the conversations you’re having at CodeAI about protecting students’ critical thinking?
It’s hard to have that conversation unless you understand the technology. You start with education around the technology.
My analogy is: We dropped a car in the middle of the classroom, gave a bunch of students keys, didn’t teach about how the car worked or what the rules of the road were about acceleration or braking, and said “drive the car.” What do we expect to happen? We’re going to have accidents, we’re going to have people going all over the place. That’s the world we’re in right now, and the outcome could be cognitive offloading, cheating, lost critical thinking, all of these things that affect students’ learning.
My take is: Let’s teach drivers. Let’s show them how a car works. Let’s explain to them how the rules of the road work. Let’s give them a framing that goes into what the car is, and then give them the car. You’ll have a better set of drivers out there.
That’s why we’re very focused on what we feel is a moral imperative to teach students about the science of these products and these technologies. If we do that, it’s one step in addressing the broader question of AI in society and education.
What does this rebranding say about what CodeAI thinks the future of computer science education should look like?
AI is like computer science’s most innovative creation. AI is born out of computer science. We wouldn’t have AI if we didn’t have the foundation of computers. Computer science continues to be foundationally an important digital science to teach about, but how we teach and what we teach is going to change.
When I grew up, my computer science curriculum taught me about memory management. Today’s software engineer and computer science program does not teach about it. They don’t need to. How we teach about computer science will change. The specifics of what we teach will change. They should change.
We will continue to teach coding. At what level, at what depth, for how long, at what intensity—that will change, but coding teaches students some fundamental practical skills. When you give a student the tool of coding, and you point them at an open blank canvas and say, solve a problem that’s relevant to yourself using technology: there’s an agency component to that, there’s an iteration component to that, there’s a perseverance component to that, and there’s a successful, confidence-building outcome from creating that thing.
