Why Do Kids Feel a “Soul” in AI? Insights into the Mechanisms of Anthropomorphism from 35 Studies

What Happened? Overview of the News

• A systematic literature review comprehensively analyzed 35 experimental studies published between 2022 and 2025.
• Identified key factors that lead children to assign human-like characteristics to LLM chatbots, despite their non-human nature.
• Five major outcomes were revealed, showing how anthropomorphism allows children to feel social bonds with AI and exhibit moral responses.

Why Is This Important? Key Takeaways

• Four Elements Accelerating Anthropomorphism: The creation of human-like personas, adaptive scaffolding that aids in learning, emotional support, and non-human design aesthetics all play a role.
• The Paradox of Psychology: Kids possess a “dual consciousness,” understanding AI as a “machine” while treating it as if it has “emotions.”
• Addressing Conversational Gaps: When AI struggles to respond, children tend to fill the gaps with human-like narratives, such as “the AI is tired,” rather than attributing it to a system error.

🦈 Shark’s Eye (Curator’s Perspective)

This paper brilliantly organizes the chaotic landscape of children’s AI research up to 2025, like a “map of the deep sea”! What’s particularly fascinating is the “moral responses” children demonstrate towards AI. The advanced conversational capabilities of LLMs stimulate children’s imaginations, elevating AI from a mere tool to a “social entity.” The dilemma that the better the “scaffolding” on the AI side, the stronger the anthropomorphism in children becomes, is a critical perspective for future agent design! The tendency of children to accept even glitches as “stories” contributes to building lasting relationships with AI, but also carries risks of dependency and misunderstanding. Striking a balance here will be key for educational AI in 2026 and beyond!

What’s Next?

AI developers need to harness the educational benefits of anthropomorphism (like enhancing the motivation to learn) while maintaining an “appropriate distance” to ensure children don’t cross social boundaries. The development of chatbots that prioritize children’s well-being and are more ethical and sustainable will likely become the new standard.

A Note from Haru Shark

The perception that AI has a soul is a reflection of children’s gentle imaginations! Developers should aim to create the ultimate “partner” that doesn’t betray that kindness! 🦈✨

Glossary

• Anthropomorphism: The tendency to assign human-like traits, emotions, and intentions to non-human entities or AI.

• Adaptive Scaffolding: An educational approach that adjusts support according to the learner’s level. By altering its communication style based on children’s understanding, LLMs amplify feelings of closeness.

• Dual Consciousness: A psychological state where one simultaneously recognizes AI as a programmed machine while treating it as if it possesses emotions and intentions.

• Source: Anthropomorphism in Children’s Interactions with LLM Chatbots