Helping Kids Recognize AI-Generated Videos
New research with 8- and 9-year-olds found that children who understood what AI is were better at spotting AI-made videos, though fewer than half identified every video correctly. The article suggests parents teach children to question unlikely scenes and look for visual or audio inconsistencies.
In an under-review study, eight- and nine-year-olds performed above chance at distinguishing AI-made videos, yet only 46.5% got every item right. Those who said they understood AI—54.7% of participants—recognized more clips correctly than peers who were less familiar.
The easiest clip for children to reject showed babies riding animals in a fashion show; many reasoned it was impossible. Some spotted odd lighting, skin, or hair, but few mentioned an on-screen AI disclosure. The authors recommend early, simple explanations and familiar examples such as voice assistants, recommendation systems, chatbots, and image generators.
As synthetic video becomes more realistic, children and families may face greater confusion about what they see online. Better AI literacy could help young viewers question implausible scenes and notice inconsistencies, but the study's low full-accuracy rate suggests even informed children remain vulnerable. Parents, educators, and platforms may need to support media literacy early, while children's trust in online content and their ability to navigate social feeds could be affected.