Google may have held back making its Bard chatbot into a publicly available product because it didn't think the technology was ready yet, according to Alphabet Chairman John Hennessy.
In an interview with CNBC, Hennessy said Bard is a "great piece of technology" to serve as a demonstration vehicle right now. He believes generative AI models are one or two years away from becoming "truly useful" tools for the general public.
- Google may have hesitated to release Bard publicly because of concerns that it would generate incorrect or harmful information, Hennessy said.
- "You don't want to put a system out that either says wrong things or sometimes says toxic things," he said.
- AI models that generate language "are still in the early days — figuring out how to bring them into a product stream and do it in a way that's sensitive to correctness, as well as issues like toxicity," Hennessey told CNBC.
- Last week, Google announced Bard, a conversational AI service, a day after Microsoft unveiled its AI-powered Bing service.
- That same day, it was found that Bard made an error, which caused Google to lose $100B in market value on Wednesday.
- The setback follows an all-hands meeting in December when Google executives said the company was taking a conservative approach to deploying AI tech in its web search.
- While the technology "strikes a need that people seem to have," language models still have many issues to address and "can make stuff up," Google AI chief Jeff Dean said.
