Children’s Understanding of Technological Omniscience

• In Event: AI in Early Childhood: Impacts on STEM and Language Learning

Abstract

Objective
Children learn from a variety of pedagogical sources: from teachers to Google searches conducted. However, neither human nor technological sources are infallible. Information about the world changes as news updates or time passes. As children’s learning environments become increasingly digital, it is vital to understand whether children believe that pedagogical informants will always be able to answer their questions.

Theoretical Framework
Some young children ascribe knowledge to a familiar informant (e.g., their mother) that they could not reasonably have (e.g., Corriveau & Harris, 2008). Young children may therefore believe teachers to be omniscient when they are not. According to the New Ontological Category (NOC) hypothesis (e.g., Kahn et al., 2012), children may ascribe non-artifact characteristics (e.g., omniscience) to technology. In recent studies where children explained why they believed Google could answer different kinds of questions, some children answered that Google knew everything or almost everything (Girouard-Hallam & Danovitch, 2022). Children may therefore increasingly believe with age that Internet-based informants can answer every, or nearly every, question.

Methods
In this study, 120 4- to 8-year-old children (59 girls, 61 boys; 75% White, 93% Non-Hispanic), were asked about whether they believed teachers and three technological sources: Google Search, the Internet at large, and a novel search engine, called Anu, could tell them the answers to every kind of question. Children were asked about their familiarity with each of these informants.

Results
Children were overwhelmingly familiar with teachers (94%), Google (84%), and the Internet (81%). Logistic regressions showed that familiarity with each of the informants was significantly positively related to age (ps<.001), suggesting that it was mostly younger children in the sample who were unfamiliar the three informants. Most children also correctly identified that they had not heard of Anu (90%).
Fifty-one percent of children said that Google was omniscient, 50% said the Internet was omniscient, 38% said that Anu was omniscient, and 23% said that a teacher was omniscient. Logistic regressions revealed that all four of these judgments were negatively related to children’s age (ps<.001; Figure 1).
A multilevel model analysis revealed a significant main effect of informant, driven by less frequent attributions of omniscience to Anu (β=-1.27, p=.002, d=-.33) and the teacher (β=-3.29, p<.001, d=-.67) than to Google or the Internet. Main effects were subsumed by a significant two-way Age and Informant interaction, driven by differences between the technological informants and the teacher (ps<.001; Table 1) and by age-related differences in children’s responses about the Internet and Google (β=-0.80, p=0.047, d=-.21; Figure 2). Younger children more frequently attributed omniscience to Google than older children did. Children’s intuitions about the omniscience of the Internet stayed more stable across age.

Scholarly Significance
Children think that technology is more capable of omniscience than teachers. Children are also less likely to evaluate novel technological informants as omniscient than familiar ones (i.e., familiarity appears to matter). Development also matters, with younger children more frequently saying that Google knew everything than that the Internet did and older children expressing the opposite intuition.

Authors

• Lauren Girouard-Hallam, University of Michigan

• Judith H Danovitch, University of Louisville