Researchers at New York University trained an AI to learn language by mimicking the way babies do, using recorded footage from a child's life.

The AI system, named the Child's View for Contrastive Learning (CVCL) model, successfully linked words to visual objects despite limited data, equivalent to one percent of the child's waking time.

The study, detailed in the journal Science, indicates AI can learn and apply words from everyday life, offering insights into early language acquisition.

Findings reveal that AI can learn a significant number of words and their application to various pictures, similar to language learning in children, utilizing contrastive learning and neural networks.