Answers are grounded in this paper and cite their sources.
Background While children interact with AI applications, they are not often given opportunities to develop a sociotechnical understanding of machine learning systems. Research is needed to understand how children make sense of AI, including algorithmic bias. We draw on sociocultural perspectives to examine how two children, aged 10 and 11, used symbolic and material tools to critique issues of bias and create personally valuable machine learning designs.Methods The research design is a descriptive single case study with embedded units. After purposeful sampling of the data, we inductively coded interviews, discussions, and digital artifacts to examine how children understood machine learning concepts and critically engaged with algorithmic bias.Findings Findings suggest using dialogue alongside model design helps children connect concepts to personal experiences and think critically about AI. Guided discussions deepened their understanding of algorithmic bias and its real-world impacts, fostering visions for more equitable technologies. However, both children developed their understanding in distinct ways.Contribution We conclude that a sociocultural approach promoting open dialogic spaces, tool appropriation, and the personal creation of machine learning applications fosters children’s understanding of sociotechnical machine learning concepts through multiple pathways, ultimately empowering them to emerge as thoughtful, responsible AI technology users and creators.
No takes yet. Share an insight, caveat, or question.
Adisa et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: