Predicting the age-appropriateness of children’s animation: Introducing The Animating Minds corpus and computational pipeline

Children’s screen time is increasing, yet much of what children watch online is neither age-appropriate nor designed for developing minds. Traditional children’s television draws on craft knowledge that aligns audiovisual stimulation with cognitive constraints, but the shift to algorithmically delivered online content has outpaced existing rating systems. The UKRI Animating Minds project addresses this gap by combining developmental science, creative practice and AI to predict the impact of animated content on young viewers. We compiled a corpus of 979 animations from UK broadcast and streaming services and extracted visual, linguistic, and narrative features to train a classifier predicting expert-rated age appropriateness. A stepwise logistic regression identified key predictors (language tone, colourfulness, shot density, frequency of visual and long words, and motion) which classified age- appropriateness of content with 77% accuracy. This model objectively confirms creative intuitions about how to design animations for pre-schoolers.