Science, Learning a New Language? ChatGpt's Fastest Young Minds

Even the smartest machines can’t compete with young minds when it comes to language learning. Researchers share new findings on how children are ahead of AI and why this matters. If a human could learn a language as fast as ChatGPT, it would take 92,000 years. While machines can process huge data sets at lightning speed, when it comes to acquiring natural language, children leave AI in the dust. A new framework recently published in the journal Trends in Cognitive Sciences by Professor Caroline Rowland of the Max Planck Institute for Psycholinguistics , in collaboration with colleagues at the ESRC LuCiD Centre in the UK , presents a new framework to explain how children achieve this remarkable feat.

Scientists can now observe, in unprecedented detail, how infants interact with their caregivers and their environment, thanks to recent advances in research tools such as head-mounted eye tracking and AI-powered speech recognition. However, despite the rapid growth in data collection methods, theoretical models explaining how this information translates into fluent speech have lagged behind. The new framework fills this gap. Synthesizing a wide range of data from computational science, linguistics, neuroscience, and psychology, the research team proposes that the key to understanding how infants learn language so much faster than AI lies not in the amount of information they receive, but in how they learn from it.

Unlike machines that learn primarily and passively from written texts, children acquire language through an active and evolving development , driven by their growing social, cognitive, and motor skills. Children use all their senses—sight, hearing, smell, listening, and touch—to make sense of the world and develop their language skills. This world provides them with rich, coordinated signals from multiple senses, providing them with diverse and synchronized cues that help them understand how language works. And children don’t just wait for language to come to them; they actively explore their environment, continually creating new learning opportunities. “AI systems process data, but children actually live it,” Rowland says. “Their learning is embodied, interactive, and deeply rooted in social and sensory contexts. They seek out experiences and dynamically adapt their learning accordingly: they explore objects with their hands and mouths, crawl toward new and exciting toys, or point to objects they find interesting. This is what allows them to master language so quickly.” These insights not only reshape our understanding of infant development, but also have far-reaching implications for research on artificial intelligence, adult language processing, and even the evolution of human language itself. “AI researchers could learn a lot from babies,” Rowland says. “If we want machines to learn language as well as humans, we may need to rethink how we design them, from scratch.”