Get Started
Home
Topics
Search
Library
Research questionHow can LLMs maintain age-appropriate safety for children ages 7–11 across languages and multi-turn conversations?General safety checks may miss whether an answer is appropriate for a child's developmental stage. Assessment is further complicated by indirect age cues, language and cultural differences, and safety changes over several turns.
AI
Alignment & Safety
Evaluation & Benchmarks
Natural Language Processing
Latest papersRecent research connected to this question, newest first.The Age of Curiosity Meets the Age of AI: Benchmarking Child Safety in Large Language ModelsThe evidence concerns child-facing LLM safety evaluated with realistic single-turn queries and multi-turn child-actor simulations, using a developmental-psychology-based LLM-as-a-Judge rubric. It covers ten query categories, no-cue, implicit-cue, and explicit-age conditions, and cross-lingual and country-context comparisons; reported results include the supplied KIDBench, KIDGuardLlama, and KIDLlama resources.research paper · Sep 2, 2026
Related questions
How should LLM safety be assessed when jailbreak vulnerability varies by language and persuasive phrasing?How can LLMs adapt to hybrid cultural influences and shifting communication preferences during multicultural dialogue?How can multi-turn LLM tutors personalize progressive guidance while preserving answer correctness?How should LLM safety be evaluated when harmful prompts vary in implicitness and sophistication?