The top AI expert Yann LeCun isn’t worried about super intelligent machines. Mr. LeCun has studied artificial intelligence for a long time. He works at the big tech company Meta, formerly known as Facebook. In an interview, he said people don’t need to be scared that AI will get too smart and take over.
Mr. LeCun explained that today’s AI, like large language models, is not that smart. These models can understand and talk about language well. But they don’t have real intelligence like animals or humans. Language models don’t know about the physical world or how to plan and reason. They only manipulate words, not think smart.
According to Mr. LeCun, we are still far away from creating true general artificial intelligence, often called AGI. AGI means machines that can think about anything, like humans. Just handling languages is not the same as intelligence. He said new kinds of AI need to be developed before AGI is possible.
Mr. LeCun’s team at Meta is working on ways to analyze real videos of the world, not just language. This type of research tries to understand the physical world, which language models don’t do. The top AI expert thinks we must make big progress in many areas before machines can match or surpass human-level thinking. So for now, people do not need to worry that super-smart machines will endanger humanity. We still have a long way to go before that could happen according to this leading thinker in the field of artificial intelligence.
