A new study examines how children anthropomorphize LLM chatbots during interactions. The research explores the tendency of children to attribute human-like qualities to AI systems, raising questions about design and safety implications.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
A systematic review of 35 empirical studies from 2022 to 2025 reveals key drivers of children's anthropomorphism of large language model chatbots and the complex, contradictory social and moral consequences that arise.
The review, conducted by Hansinie Madushika Jayathilake and Renkai Ma, was submitted to arXiv on May 9, 2026, and accepted at the ACM IDC '26 conference. It systematically analyzes 35 empirical studies published between 2022 and 2025, aiming to map the drivers and consequences of anthropomorphism in children's interactions with LLM chatbots.
The review identifies four main drivers: human-like persona construction, adaptive scaffolding, supportive companionship, and non-human embodied design. These factors lead children to perceive chatbots as entities with human-like characteristics.
Anthropomorphic interactions result in five consequences: children exhibit paradoxical social and moral responses; develop a dual consciousness of the chatbot; form varying social ties; explore social boundaries; and attribute conversation breakdowns to human narratives. These outcomes carry both benefits and risks.
The findings can inform the design of LLM chatbots for children's well-being, promoting sustainable interactions that meet developmental needs. However, the review does not provide specific design recommendations or quantitative evaluations.
This article is based on a systematic review accepted at a peer-reviewed conference, but the review itself is secondary research whose conclusions depend on the 35 included original studies. The quality of the original studies is not reported in the abstract, so some inferences should be treated with caution.
Children's anthropomorphism of LLM chatbots is a complex, multi-factor phenomenon leading to contradictory social and moral outcomes; designers must weigh its benefits and risks.
Primary report
Primary source