Infants Outperform Artificial Intelligence in Uncovering Drivers of Human Behavior A recent study conducted by a research team comprised of psychology and data science experts has revealed that infants are better-equipped than Artificial Intelligence in understanding what motivates people’s actions. This breakthrough surface many fundamental differences between cognition and computation and has illuminated flaws in existing technologies and how AI needs to be improved to better mimic human behavior. This research team was led by Professor Roberta Golinkoff from the University of Delaware in collaboration with the University of Massachusetts, Amherst and the University of California, Irvine. They used eye-tracking technology and video tracking software to study infants and AI’s level of detection of what motivates a person’s behavior. The team compiled a set of videos of an adult interacting with two jars of different colors – one contained coins and the other contained balls – and interacted with the containers using an array of different approaches. For example, in one scenario the adult would pour coins from one container to another, while in another they would put a ball from one container into the other. The team then asked AI and infants to identify which container held coins and which contained balls. The results showed that infants aged 6 to 12 months had a significantly higher success rate at recognizing which jar contained coins compared to the AI system. In a statement released by Professor Golinkoff, she stated that "This study is one of the first to compare the abilities of infants and AI in understanding human behavior and motivation.” The authors of this research believe that the results of the study demonstrate a key difference between an infant’s robust understanding of what drives human behavior and an AI system’s weaker understanding of the same. They believe this difference is down to a difference in the cognitive abilities of infant brains and AI systems. AI systems typically use two processes to understand a situation – pattern recognition and predictive analytics. Pattern recognition determines whether a given object looks similar to any other objects, while predictive analytics works out the possible outcomes of a situation based on past data. On the other hand, human cognition is capable of much more than this, as it is capable of perceiving the environment from various angles and in various contexts. This means that when AI looks at a video of a person interacting with two containers, it can easily recognize the objects but may struggle to identify which jar the person is likely to choose. On the other hand, infants have a much more sophisticated understanding of their environment and of the motivations behind the adult’s behavior. The authors suggest that more research needs to be conducted to understand how AI can replicate the level of sophistication of infants’ understanding of their environment. They also believe that this research points to the need to focus on developing AI that is better-equipped with social cognition capabilities. In conclusion, the findings of this research indicate that infants are much better at uncovering what motivates other people’s actions than AI systems. This highlights fundamental differences between cognition and computation, exposing the shortcomings of current technologies and the need to make improvements. Moreover, this research provides pointers towards the direction of AI development, suggesting that AI systems need to be better-equipped with social cognition capabilities. As AI continues to develop, it will be interesting to see if these systems can replicate and even surpass the level of sophistication of infant brains.
https://www.lifetechnology.com/blogs/life-technology-medical-news/infants-outperform-ai-in-commonsense-psychology
Buy SuperforceX™
