HCDE impact: Helping kids think critically about AI

As artificial intelligence (AI) becomes increasingly woven into daily life, promoting AI literacy among young people is more important than ever. Researchers from the University of Washington’s Computing for Healthy Living and Learning (CHiLL) Lab, led by Human Centered Design & Engineering Professor Julie Kientz, are exploring innovative ways to help kids build the critical thinking skills they need to understand and engage with AI.

One new project, led by PhD student Aayushi Dangol, is AI Puzzlers, an interactive game designed to help children explore the limits of generative AI reasoning. Using block-based visual puzzles inspired by the Abstraction and Reasoning Corpus (ARC), AI Puzzlers highlights that while AI can do many things, it struggles with abstract reasoning in ways that humans, even young children, easily master.

“Through our research, we’ve seen that kids often have a mental model of AI as an all-knowing robot entity—it knows everything, it has the entire internet. We're trying to help kids develop a better mental model of when AI can be a helpful tool, and when it might mislead them,” said Kientz.

Recognizing that children are naturally drawn to games and puzzles, Dangol adapted ARC puzzles, originally developed to benchmark AI progress, into a web-based game. Drawing on Mayer and Moreno’s Cognitive Theory of Multimedia Learning, AI Puzzlers combines visual and verbal elements to reduce cognitive overload and support children's ability to detect AI errors. This design allows kids to visually compare their puzzle solutions to those generated by AI, encouraging meaningful engagement and critical thinking.

“Even for adults, recognizing errors in generative AI isn’t always straightforward, said Dangol. “But the visual nature of these puzzles makes AI errors easy to spot. We saw children who weren’t yet fluent readers quickly identify inconsistencies in AI-generated solutions.”

Testing AI Puzzlers: insights from kids

Dangol first deployed AI Puzzlers during the 2024 Engineering Discovery Days at UW, where over 120 K-12 students from across Washington State interacted with the system. Dangol also led two participatory design sessions with 21 children (ages 6–11), studying how kids recognize errors in AI reasoning, develop strategies for navigating these errors, and evaluate AI outputs.

Rather than viewing AI as an all-knowing machine, young participants began to see that AI, like any tool, has both strengths and weaknesses. Many children expressed surprise and amusement when AI made mistakes on puzzles they found easy, leading to discussions about "how AI thinks" and how it differs from human reasoning.

Looking ahead, the researchers envision AI Puzzlers as a valuable tool for elementary school teachers to have conversations about AI literacy in a playful and engaging way. By helping children experience firsthand how AI can make mistakes, it supports critical thinking at an early age.

Research with real-world impact

The importance of this research is underscored by broader national efforts: a White House executive order recently called for promoting AI literacy in K–12 education. The CHiLL Lab’s work directly supports this mission, helping young learners build the critical thinking skills needed to navigate a world increasingly influenced by AI.

Backed by funding from the AI Institute for Exceptional Education, funded jointly by the National Science Foundation and the Department of Education’s Institute for Educational Sciences, this research reflects the value of academic work that generates knowledge that can transfer across systems and contexts. 

“Academic research plays a crucial role in helping society keep pace with rapid technological change,” said Kientz. “Unlike industry, we're not driven by profit, we're driven by the need to understand, to educate, and to generate new knowledge that serves the public good. This kind of work helps ensure that families, educators, and policymakers have access to thoughtful, research-based insights on emerging technologies like AI.”

Aayushi Dangol will present the AI Puzzlers findings at the Interaction Design and Children (IDC) Conference in Iceland in June 2025, contributing to a growing body of research from the CHiLL Lab focused on children’s healthy engagement with technology.

Through projects like AI Puzzlers, the UW’s CHiLL Lab continues to push forward how we think about technology, education, and child development, while equipping the next generation to understand and question the role of AI in their lives.

View AI Puzzlers in action in this three-minute video: