Skip to main content
RealTime Wire
ScienceInternational

Scientists Leverage A.I. to Decode Human Cognition

A new A.I. system mimics human behavior in psychological experiments, providing insights into cognition and decision-making.

By Carl ZimmerSource: nytimes.com1 min read

In brief

An international team of researchers has developed an A.I. known as Centaur, which accurately simulates human cognitive processes, aiding the understanding of how human minds function through psychological experiments.

In a groundbreaking effort to unravel the intricacies of human cognition, scientists have unveiled a novel artificial intelligence system named Centaur. This innovative project is designed to effectively mimic human behavior in psychological experiments, offering a fresh lens through which to explore and validate existing cognitive theories. The work, recently published in the journal Nature, highlights the potential of A.I. to not only achieve task-specific capabilities, such as chess playing and language generation, but also contribute to the field of psychology.The race towards artificial general intelligence (AGI) has seen significant financial investment from tech giants like OpenAI and Meta, yet the concept remains nebulous, primarily serving as an aspirational goal rather than a clearly defined target. Current A.I. systems demonstrate remarkable proficiency in certain domains while lacking the versatility of human minds. For instance, despite being champion chess players, these A.I. entities are clueless when faced with driving tasks. They can also exhibit peculiar errors, such as allowing illegal moves in chess games.The Centaur system bridges this gap by employing A.I. to simulate human responses in psychological research, which allows scientists to test cognitive theories more effectively. Over recent years, cognitive scientists have developed comprehensive frameworks to interpret mental processes like decision-making and memory recall. Through experiments, these theories have been validated or refined based on how closely human behavior aligns with predicted models.A notable observation from these studies is that individuals generally prefer certain outcomes over risky ventures, even if the latter could yield significantly greater rewards. For instance, when presented with a guaranteed $1,000, most people would choose that option over a gamble with a potentially higher payout but inherent uncertainty.Overall, Centaur represents a significant step forward in the application of A.I. within psychology, paving the way for deeper understanding of human cognition and prompting further exploration into the capabilities and limitations of artificial intelligence in this domain.

Topics

More on this story