Living in a digital ecology: Children's selective trust in technological informants

Children learn not only from observation and interaction with their environment but also from the testimony of others. In the digital age, interactive technologies increasingly serve as information sources alongside peers, caregivers, and teachers. This study investigated whether children selectively learn from technological informants in a novel-word learning context. 180 Chinese preschoolers (M = 49.17 months, SD = 7.04 months, 51.7 % girls) were randomly assigned to one of six conditions: accurate voice assistant, inaccurate voice assistant, accurate robot, inaccurate robot, accurate human, or inaccurate human. Children first engaged in an agency beliefs interview regarding their corresponding informants followed by a classic selective trust task. We found that children selectively trusted (and distrusted) voice assistants and robots based on their past accuracy, while they demonstrated a general trust in humans, even when the humans were previously inaccurate. Agency beliefs explained variations in children's trust of voice assistants and humans. We also observed developmental differences in children's levels of trust and distrust toward both accurate and inaccurate informants. These findings suggest that 3–4-year-olds engage in both epistemic and social considerations in selective trust, and their judgments are susceptible to the agentic features of informants.