# What Do Children and Parents Want and Perceive in Conversational Agents? Towards Transparent, Trustworthy, Democratized Agents

**Authors:** Jessica Van Brummelen, Maura Kelleher, Mingyan Claire Tian, Nghi Hoang Nguyen  
**arXiv:** [2209.07862](https://arxiv.org/abs/2209.07862)  
**Date:** 2022-09-16  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2209.07862v2/

## Key points

- Historical focus on analyzing technology from a WEIRD adult perspective
- Broadening the perspective by analyzing viewpoints of children and parents from various countries regarding conversational agents
- Empowering children and parents to program their own agents through educational workshops
- Children perceived agents as more human-like, warm, and dependable compared to parents' views
- Participants trusted agents more than parents or friends for accurate information
- Children described ideal agents as more artificial than human-like, prioritizing fun features and approachable characteristics
- Implications for agent design: focus on indicators of competence and predictability, increase transparency about information sources
- Importance of considering diverse perspectives when designing technologies like conversational agents

## Summary

In their study "What Do Children and Parents Want and Perceive in Conversational Agents? Towards Transparent, Trustworthy, Democratized Agents," authors Jessica Van Brummelen, Maura Kelleher, Mingyan Claire Tian, and Nghi Hoang Nguyen address the historical focus on analyzing technology from a WEIRD (Western, Educated, Industrialized, Rich, and Democratic) adult perspective. This limited approach may result in technology that is not appropriately developed for children or individuals from non-WEIRD countries. The researchers aimed to broaden this perspective by analyzing the viewpoints of children and parents from various countries regarding conversational agents. By understanding participants' trust in agents, partner models, and their ideas of "ideal future agents," the study sought to inform better design practices for these users. Additionally, the researchers empowered children and parents to program their own agents through educational workshops to observe changes in perceptions as participants engaged with and learned about agents. The results of the study involving 49 participants revealed interesting insights. Children perceived agents as significantly more human-like, warm, and dependable compared to how parents viewed them. Participants trusted agents more than they trusted parents or friends for accurate information. Children described their ideal agents as being more artificial than human-like in contrast to parents' perspectives. Moreover, children tended to prioritize fun features, approachable/friendly characteristics, and addressing concerns through agent design more than parents did. The study also discussed potential implications for agent design based on the findings. Designers may enhance levels of trust towards agents by focusing on developing indicators of competence and predictability within the agents while also increasing transparency regarding the sources of information utilized by these conversational agents. Overall,this research sheds light on the importance of considering diverse perspectives when designing technologies like conversational agents.By incorporating insights from both children and parents across different cultural backgrounds,designers can create more transparent, trustworthy, and democratized agent models that cater to a broader range of users effectively.