Human psychology: infants outperform artificial intelligence in some aspects

More and more scientific studies highlight the progress of Artificial Intelligence, and in particular in terms of medical diagnostics, where it would do better than doctors. For human psychology, on the other hand, it seems more complicated.

Infants have indeed outperformed artificial intelligence (AI) technology in an experiment involving human psychology. In a new study, published online this February 16 in the journal Cognition (Source 1), American researchers report having observed the superiority of infants over AI in detecting what motivates the actions of others.

The research team here conducted a series of experiments with 11 month old infants and artificial intelligence, using the “Baby Intuitions Benchmark” (BIB), a suite of tasks that challenge both infants and AI to make high-level predictions regarding the motivations that lead to stock.

Via the Zoom software, the babies watched videos of animated shapes moving on the screen, like a video game. The actions of the animated shapes simulated human behavior and decision making to retrieve objects from the screen. The AI ​​was also exposed to the same video content.

Verdict: The infants recognized human motivations, even in the actions of the animated shapes. They correctly predicted that these actions were motivated by retrieving objects from the screen, a parameter quantified via babies’ gaze time.

Previous studies had already highlighted the common sense psychology » infants. They are indeed able to attribute objectives to others, to understand their intentions and preferences. “ The ability to make these predictions is fundamental to human social intelligence “, assure the researchers in a press release (Source 2), who believe that what is still lacking in AI, “ It is flexibility in recognizing different contexts and situations that guide human behavior.

Fundamental differences between man and machine

For Moira Dillon, assistant professor in the Department of Psychology at New York University and lead author of the study, these results “ highlight fundamental differences between cognition and computation »as much as they point out “ lThe shortcomings of current technologies and areas where improvements are needed for AI to more fully replicate human behavior “. Because if adults and infants can easily make reliable inferences on what motivates the actions of others”, the AI ​​seems unable to do so. At least for the moment.

Leave a Replay