Google CEO Admits They Don’t Know How Their AI Works

Google CEO Sundar Pichai admitted that neither he nor the team at Google understands how the artificial intelligence that they created works, according to The Post Millennial. His comments came during a “60 Minutes” interview

After the launch of ChatGPT, the large language model (LLM) that users can interact with and message, Google constructed its own bot, called “Bard.” But Pichai said that Bard developed “emergent properties,” a phenomenon that occurs when AI learns skills that it was not designed for. Pichai said that this was not something he “fully understands.” 

He then mentioned that they call this “black box,” which is a term describing an internal mechanism in a technological device that is unknown to the user or creator. He said that this is the “state of the art” and that over time their understanding of the technology will improve. 

There are looming questions surrounding whether society is ready for AI. Elon Musk has claimed that the technology will cause “civilization destruction,” but has decided to invest in it anyway and create a competitor to the other services that he claims are biased. He has reportedly joined other tech leaders in issuing a letter calling for a six-month pause in the “out of control” AI race. 

Pichai said that on the one hand, society was not ready for AI because there is a mismatch between the pace of technology and the pace at which humans can think and adapt. On the other hand, he said he feels “optimistic” because he’s “seen more people worried about it earlier in its life cycle.” But he emphasized that how society reacts is not for a company to decide, but the society itself. 

New developments in the space revealed that ChatGPT passed the U.S. Medical Licensing Exam (USMLE) without any human help.