AI chatbots sometimes talk nonsense seriously, Google executives warned after the debut.

CTOnews.com, February 12 (Xinhua)-- Bard, Google's new AI chat robot, gave the wrong answer to a query this week, causing the share price of its parent company Alphabet to plunge 8%, reducing its market capitalization by more than $100 billion.

Google search engine executives said in an interview with the media on Saturday that current forms of generative artificial intelligence, such as ChatGPT, may be unreliable and should beware of artificial intelligence traps in chatbots.

Prabhakar Raghavan, senior vice president and head of Google Search, said in an interview: "the kind of artificial intelligence we are talking about sometimes leads to what we call an hallucination, in which the machine gives a seemingly convincing but completely fabricated answer."

Lagawan added that one of the basic tasks of the artificial intelligence industry at present is to keep this situation to a minimum.

CTOnews.com learned that, in fact, many ChatGPT users, including Apple co-founder Steve Wozniak, complain that they often give the wrong answer.

Google developers have been under tremendous pressure since the launch of OpenAI's ChatGPT, which has become popular around the world and threatens Google's core business. "We obviously feel the urgency, but we also feel a huge responsibility," Ms. Lagawan said. "We certainly don't want to mislead the public."