[論文レビュー] Understanding Parents' Desires in Moderating Children's Interactions with GenAI Chatbots through LLM-Generated Probes

John Driscoll, Yulin Chen+4 more|arXiv (Cornell University)|Mar 4, 2026

AI in Service Interactions被引用数 0

ひとこと要約

The paper investigates how parents want to moderate child–GenAI chatbot interactions, using LLM-generated scenarios to elicit concerns, moderation preferences, and transparency needs from 24 parents.

ABSTRACT

This paper studies how parents want to moderate children's interactions with Generative AI chatbots, with the goal of informing the design of future GenAI parental control tools. We first used an LLM to generate synthetic child-GenAI chatbot interaction scenarios and worked with four parents to validate their realism. From this dataset, we carefully selected 12 diverse examples that evoked varying levels of concern and were rated the most realistic. Each example included a prompt and a GenAI chatbot response. We presented these to parents (N=24) and asked whether they found them concerning, why, and how they would prefer the responses to be modified and communicated. Our findings reveal three key insights: (1) parents express concern about interactions that current GenAI chatbot parental controls neglect; (2) parents want fine-grained transparency and moderation at the conversation level; and (3) parents need personalized controls that adapt to their desired strategies and children's ages.

研究の動機と目的

• Identify factors that trigger parental concern in Child–GenAI Chatbot interactions.
• Understand the moderation strategies parents want for age-appropriate, safe, and values-aligned interactions.
• Characterize parents' transparency preferences and how they want to be involved.

提案手法

• Generate synthetic Child–GenAI Chatbot interaction scenarios with an LLM and validate realism with parents.
• Select a diverse set of scenarios through iterative filtering and parent validation.
• Conduct semi-structured interviews with 24 parents using 12 selected scenarios to elicit concerns, moderation needs, and transparency preferences.
• Apply thematic analysis to interview transcripts to derive high-level themes.

Figure 1 . We ➀ presented 12 simulated GenAI Chatbot scenarios to 24 parents and ➁ asked whether they found them concerning, why, and how they would prefer to modify the responses and be informed. Our analysis identifies the factors parents found concerning which includes ➂ types of questionable res

実験結果

リサーチクエスチョン

• RQ1 RQ1: What factors trigger parents' concern in Child–GenAI Chatbot interactions?
• RQ2 RQ2: How do parents want to moderate their children’s interactions with GenAI Chatbots?
• RQ3 RQ3: What are parents’ transparency preferences around their child’s interactions with a GenAI Chatbot?

主な発見

• Parents’ concerns center on two groups: questionable GenAI chatbot responses and questionable child prompts.
• Parents desire fine-grained, conversation-level moderation with goals including age-appropriate language, corrected misunderstandings, and deferral to human support.
• Transparency preferences cluster around involvement and data-content access, with many parents wanting real-time alerts for flagged activity.
• Parents seek personalized, developmentally appropriate moderation that adapts to their child’s age and family values.
• Interviews suggest the chatbot should function as a mediator that supports the parent’s ongoing role in safety and development.

Figure 2 . Our study procedure. We design, generate, select, and present realistic scenarios to parents as probes in semi-structured interviews, then analyze parents’ responses.