Language models are inferior to dogs in intelligence, says Meta’s chief scientist
Large artificial intelligence language models are not truly intelligent because they cannot understand, interact or make sense of reality and rely on language training to achieve results. This was stated by Yann LeCun, Meta’s chief researcher for artificial intelligence, writes ZDNET.
According to him, large language models are not even as intelligent as dogs. He explained that true intelligence transcends language, because most human knowledge has little to do with language. Programs like ChatGPT lack emotion, creativity, sensuality and consciousness. That is, the cornerstones of human intelligence.
Yann LeCun believes that one day artificial beings will become more intelligent than humans, and when that happens, they will have to be “controlled and essentially subordinated to humans.” He also believes that people’s fear of artificial beings wanting to take over the world is unfounded, as there is no correlation between being intelligent and wanting to take over the world.
We will remind that Great Britain this fall will host the first major Global Summit on artificial intelligence security. At the event, it is planned to consider the risks associated with artificial intelligence. It will also discuss how they can be mitigated through internationally coordinated action.