Ask Me Anything: Exploring children's attitudes toward an age-tailored AI-powered chatbot

TL;DR

Ask Me Anything investigates children's attitudes and trust toward an age tailored AI chatbot that answers only topic specific questions in astronomy, sneakers, and dinosaurs. The study deploys AMA in a K-8 setting, collecting transcripts, logs, surveys, and interviews to examine engagement, trust development, and anthropomorphism. Findings indicate high openness and trust with three main conversational themes and active reliability testing, underscoring the need for digital safety and AI literacy in schooling. The work contributes empirical insights into child interactions with constrained AI agents and informs design of child friendly AI education tools.

Abstract

Conversational agents, such as chatbots, have increasingly found their way into many dimensions of our lives, including entertainment and education. In this exploratory study we built a child-friendly chatbot, "Ask Me Anything" (AMA), and investigated children's attitudes and trust toward AI-driven conversational agents. To prompt targeted questioning from students and drive engagement, AMA is a specialized chatbot that answers only topic--specific questions in three areas--astronomy, sneakers and shoes, and dinosaurs. We tested AMA with 63 students in a K-8 public school in the Northeast USA. Students worked in small groups, interacted with our tool for three to ten minutes, and completed a post survey. We identified three key themes that emerged from student conversational interactions with AMA: expressing wonder, surprise, and curiosity; building trust and developing confidence; and building relationships and anthropomorphizing. Also, we observed a broad attitude of openness and comfort. Students trusted the chatbot responses in general, indicating a high level of trust in and reliance on AI as a source of information. They described AMA as "knowledgeable," "smart," and that they could "trust it." To confirm their perception of reliability, some students tested the chatbot with questions to which they knew the answers. This behavior illustrated a fundamental aspect of children's cognitive development: the process of actively evaluating the credibility of sources. Our work extends and contributes to the existing body of literature that explores children's interactions with conversational agents.

Figures (10)

• Figure 1: Ask Me Anything user interface. The user interface consists of two sides. On the left side, users can enter their name, grade, age range, and select a topic from the drop-down menu. On the right side, at the top, users can write their question and click the submit button. The AI's answer will then be displayed on the right side, below the submit button, in the answer box Robot and sneaker images are licenced for academic use from canva.com.
• Figure 2: An example of the ChatGPT 3.5 turbo model going off-topic after repeated questioning.
• Figure 3: An 8th grade student, James (a pseudonym), asked AMA about itself. Our deployed system used text-davinci-002. The app stayed on topic, responding only questions about "sneakers and shoes." The student responded with a question about this topic and AMA answered accordingly.
• Figure 4: Student engagement by questions asked. Most students posed one to three questions (37 out of 63 students) and 13 students asked 4 to 6 questions. A total of 13 students asked seven or more questions, including a first grader who asked 21 questions, the highest number in the study.
• Figure 5: Topic popularity by category. Among Students- topic "dinosaurs" emerged as a favored subject with 33 students initiating or switching mode to this topic, , closely followed by"sneakers and shoes" with 29 students expressing interest through questions or topic changes