AI is unlikely to achieve human-like cognition unless it’s connected to the real world via robots, according to a study
According to research by the University of Sheffield, the best way to achieve human-like cognition in AI is to connect artificial intelligence systems with the real world via robots and design them using principles of evolution.
Professor Tony Prescott, Dr. Stuart Wilson, and the Department of Computer Science at the University of California, Berkeley, wrote in a recent paper that AI systems, even if they are disembodied, will not be able to mimic real brain functions, no matter how big their neural networks, or datasets to train them, may become.
Current AI systems such as ChatGPT use large neural networks in order to solve complex problems such as the generation of understandable written text. These networks train AI to process data using a method inspired by the brain, and to learn from mistakes to become more accurate.
Source:
https://techxplore.com/news/2023-06-ai-gain-human-like-cognition-real.html
