# LLMs and Childhood Safety: Identifying Risks and Proposing a Protection Framework for Safe Child-LLM Interaction

arXiv: https://arxiv.org/abs/2502.11242 · HTML: https://mlsanity.com/paper/2502.11242/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 13 of its 110 references that are in the map.
- First posted on arXiv 16 Feb 2025.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 13 of 13)

- [How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs](https://mlsanity.com/paper/2401.06373/) (2024)
- [Toxicity in ChatGPT: Analyzing Persona-assigned Language Models](https://mlsanity.com/paper/2304.05335/) (2023)
- [The Ethics of Advanced AI Assistants](https://mlsanity.com/paper/2404.16244/) (2024)
- [Improving Human-AI Partnerships in Child Welfare: Understanding Worker Practices, Challenges, and Desires for Algorithmic Decision Support](https://mlsanity.com/paper/2204.02310/) (2022)
- [A Piece of Theatre: Investigating How Teachers Design LLM Chatbots to Assist Adolescent Cyberbullying Education](https://mlsanity.com/paper/2402.17456/) (2024)
- [Can generative AI and ChatGPT outperform humans on cognitive-demanding problem-solving tasks in science?](https://mlsanity.com/paper/2401.15081/) (2024)
- [Towards social generative AI for education: theory, practices and ethics](https://mlsanity.com/paper/2306.10063/) (2023)
- [Multimodal Chain-of-Thought Reasoning via ChatGPT to Protect Children from Age-Inappropriate Apps](https://mlsanity.com/paper/2407.06309/) (2024)
- [Distributed agency in second language learning and teaching through generative AI](https://mlsanity.com/paper/2403.20216/) (2024)
- [Helpful or Harmful? Exploring the Efficacy of Large Language Models for Online Grooming Prevention](https://mlsanity.com/paper/2403.09795/) (2024)
- [The Future of Child Development in the AI Era. Cross-Disciplinary Perspectives Between AI and Child Development Experts](https://mlsanity.com/paper/2405.19275/) (2024)
- [App Planner: Utilizing Generative AI in K-12 Mobile App Development Education](https://mlsanity.com/paper/2401.15182/) (2024)
- [Malak: AI-based multilingual personal assistant to combat misinformation and generative AI safety issues](https://mlsanity.com/paper/2410.11856/) (2024)

## Built on by (most cited first, 7 of 7)

- [Are Rules Meant to be Broken? Understanding Multilingual Moral Reasoning as a Computational Pipeline with UniMoral](https://mlsanity.com/paper/2502.14083/) (2025)
- [MinorBench: A hand-built benchmark for content-based risks for children](https://mlsanity.com/paper/2503.10242/) (2025)
- [Empowering Children to Create AI-Enabled Augmented Reality Experiences](https://mlsanity.com/paper/2508.08467/) (2025)
- [AutiHero: Engaging Parents in Creating Personalized, Multi-path Social Narratives for Autistic Children](https://mlsanity.com/paper/2509.17608/) (2025)
- [The Fake Friend Dilemma: Relational Trust and the Political Economy of Conversational AI](https://mlsanity.com/paper/2601.03222/) (2026)
- [The Age of Curiosity Meets the Age of AI: Benchmarking Child Safety in Large Language Models](https://mlsanity.com/paper/2605.25510/) (2026)
- [Understanding Parents' Complex Views of AI for Children's Pretend Play](https://mlsanity.com/paper/2609.39906/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) are increasingly embedded in child-facing contexts such as education, companionship, creative tools, but their deployment raises safety, privacy, developmental, and security risks. We conduct a systematic literature review of child-LLM interaction risks and organize findings into a structured map that separates (i) parent-reported concerns, (ii) empirically documented harms, and (iii) gaps between perceived and observed risk. Moving beyond descriptive listing, we compare how different evidence streams in surveys, incident reports, youth interaction logs, and governance guidance operationalize "harm," where they conflict, and what mitigations they imply. Based on this synthesis, we propose a protection framework that couples child-specific content safety and developmental sensitivity with security-grade controls for adversarial misuse, including prompt injection and multimodal jailbreak pathways. The framework specifies measurable evaluation targets (e.g., harmful-content avoidance, age-calibrated readability, bias parity checks, prompt-injection robustness, and monitoring transparency) to support developers, educators, and policymakers in assessing and improving child-safe LLM deployments.