How LEGO Is Promoting K–8 Computer Science

Building From the Ground Up With Bricks and Machine Learning

The curriculum consists of kits that come in stackable boxes and are available in three grade bands (Kindergarten through second grade, third grade through fifth grade, and sixth grade through eighth grade). Each kit, which is meant to be used by four students working together, contains LEGO bricks, interactive hardware, one or two connection cards, a charging cable and building instructions.

“We’ve integrated motors and sensors into LEGO elements, so there are four tech-enabled components in the system,” says Sliwinski. Students can connect those elements to explore computational thinking concepts, and then connect them to a computer for more sophisticated programming.

Every kit includes a scope and sequence with 30 inquiry-based, curriculum-aligned lessons that can integrate seamlessly into existing programs or provide stand-alone computer science and AI curriculum coverage.

RELATED: AI literacy for K–12 students is top of mind among educators.

The software has two halves: “There’s a teacher-facing half, which we call the teacher portal, which has lessons, curriculum, teacher preparation notes, teacher presentations — everything you need to go from zero to running your lesson,” Sliwinski says.

The student-facing half of the software is the actual coding experience. For younger children, there is an icon-based programming language that’s similar to ScratchJr. For older kids who can read, there’s word-based programming grammar that’s similar to Scratch. (Before coming to LEGO Education, Sliwinski was co-director of Scratch.)

AI tools are built into both the teacher and student-facing portions. Students can explore pretrained machine learning models and even train their own models to interact with different LEGO hardware.

Focusing on AI Fundamentals, Not Just Tools

Sliwinski argues that the tech industry has become so fixated on the next big AI model or tool that it prioritizes teaching students how to use those products rather than helping them truly understand how the technology works.

“What we’re focused on here is that level one down — getting to that foundational understanding of AI and of AI concepts,” he says.

“Our perspective is that we shouldn’t be using AI on kids. We should be putting AI in kids’ hands,” he adds. “Yes, we need to introduce them to these tools, but we also need to put them in control. That idea was really the genesis of the product.”

Emphasizing Collaboration and Students' Interests

Getting students working in small groups was another priority for the curriculum, he says.

“The kids are actively working together. They build with each other, talk to each other, learn from each other, problem-solve together, fight with each other,” he says. “Sometimes they have to navigate tricky conversations together.”

The lessons are designed to directly connect with what children are interested in. Sliwinski says one of his favorite lessons for grades three to five enables students to tap into their love of TikTok- and Fortnite-style dance trends by training a machine learning model to recognize different moves and then teaching a robot to copy them. It’s all about getting “connected to what kids care about.”

Breaking out of that mold, he says, allows teachers to cover a wide range of computer science ideas instead of having “kids do algorithms again and again.”

“It's not that there's anything wrong with algorithms. That’s a really powerful skill to develop or knowledge to have. But there's so much more to computer science than just algorithms,” he adds. “And so, when the robot can be a dancer or a drawing machine or a CNC machine or a monster that's chasing you around the classroom, all of a sudden we're able to hit a much broader set of computational thinking concepts as well.”

SUBSCRIBE: Sign up to get the latest EdTech content delivered to your inbox weekly.

Screen-Smart Strategy and Strong Privacy Guarantees

With the youngest students, he says, the team is “very intentional about delaying screens” because devices add complexity and tend to dominate early learners’ attention. In kindergarten through second grade, students start off-screen to build core computational thinking skills first, and only “once those foundations are solid” do they introduce screens, so that “the tech supports the learning instead of distracting from it.”

From the start, he says, the team also drew clear “red lines” around safety, privacy and security — nonnegotiable boundaries tied to LEGO’s values. In practice, that means “there are no student accounts and no data leaving the classroom.”

When students train a machine learning model, “all of the inference runs locally on their device,” so kids can train their own models even on 10-year-old Chromebooks, and “none of that data is ever sent to a third party, including us.”

In the end, LEGO Education isn’t just trying to make AI less intimidating. It’s also trying to make sure today’s students don’t remain passive users of AI but grow into the active builders who decide what the technology becomes next, he says.

“Children are the architects of the future,” Sliwinski says. They don’t just need to be able to use these tools. They need to know how to build things with them and how they work. “They’re going to build the next generation that comes after them.”