Today, Artificial Intelligence is everywhere. From self-driving cars to speech recognition, artificial intelligence is an integral aspect of our daily life. A recent study revealed, however, that infants outperform AI in crucial psychological skills.
Recent research from New York University and published in the journal Cognition reveals that infants can perceive the motive behind a person’s gestures better than artificial intelligence. This study emphasizes the significance of upgrading current technology and identifying AI’s limitations by demonstrating the unique distinction between cognition and computing.
Moira Dillon, Ph.D., an assistant professor in the Department of Psychology at New York University and the paper’s principal author, stated, “Adults and even toddlers may make reliable judgments about what motivates other people’s behavior.” “Present AI finds it difficult to draw these conclusions.”
She says, “The new idea of placing infants and AI head-to-head on the same tasks enables researchers to better define infants’ intuitive knowledge about other people and to suggest strategies to incorporate this knowledge into AI.”
Babies are fascinated by people, as seen by their fixation on and perception of humans. They can also converse with others and comprehend the most fundamental emotions of others. Infants develop human social intelligence as a result of their abilities, which include the formation of objectives and the formation of preferences.
The study investigated the distinction between infants and artificial intelligence by examining 84 11-month-old infants. The team observed infants’ responses to the’state-of-the-art learning-driven neural-network model’ by comparing them to AIs. The team utilized the “Baby Intuitions Benchmark” (BIB), which consists of six tests that investigate commonsense psychology.
BIB was designed to evaluate infant and machine intelligence, enabling a comparison of infant and machine performance and, most significantly, providing a measurable foundation for the development of humanistic AI. On Zoom, infants viewed videos including basic, animated shapes that rolled across the screen. The simulation of human behavior and decision-making involved the retrieval of on-screen objects and other movements.
Similarly, the team developed, trained, and evaluated learning-driven neural network models, or AI machines, which aid computers in recognizing patterns and replicating human intellect. The team discovered that infants could recognize human-like intentions in basic actions and animated shapes. They were tasked with recognizing the retrieval of identical objects on the screen despite the ever-changing environment. Longer staring at moving things was indicative of infant recognition.
On the contrary, AI tools failed to demonstrate any recognition. This capacity to consider actions and data appear to be unique to humans at present. It allows us to interact with others and collaborate.
Dillon concluded, “A human child’s basic knowledge is limited, abstract, and reflective of our evolutionary heritage, yet it is adaptable to whatever setting or culture in which the infant may live and learn.”
