oogle’s much-touted new AI chatbot tool Bard, which has yet to be publicly released, is already being chastised for an incorrect response it produced in a demo this week.
A user asks Bard in the demo, which Google posted on Twitter, “What new discoveries from the James Webb Space Telescope can I tell my 9-year-old about?” Bard responds with a series of bullet points, one of which reads, “JWST took the very first images of a planet outside of our own solar system.”
According to NASA, the first image of an exoplanet — or any planet beyond our solar system — was taken nearly two decades ago, in 2004, by the European Southern Observatory’s Very Large Telescope.
Following the inaccurate response from Bard, shares in Google’s parent company Alphabet fell 7.7% Wednesday, wiping $100 billion off its market value.
Bard’s gaffe highlights Google’s challenge in integrating the same AI technology that powers Microsoft-backed ChatGPT into its core search engine. In order to keep up with what some believe will be a radical shift in how people search online caused by conversational AI, Google now risks undermining its search engine’s reputation for surfacing reliable information.
Bard, like ChatGPT, is based on a large language model that is trained on massive amounts of data online to generate compelling responses to user prompts. Long have experts warned that these tools have the potential to disseminate inaccurate information.
In an apparent attempt to address that concern, Google previously stated that Bard would be made available to “trusted testers” this week, with plans to make it public in the coming weeks.
“This highlights the importance of a rigorous testing process, which we’re launching this week with our Trusted Tester programme,” a Google spokesperson said in a statement to CNN about the factual error on Wednesday. “We’ll combine external feedback with our own internal testing to ensure that Bard’s responses meet a high standard for quality, safety, and grounding in real-world data.”
Alphabet shares fell as much as 8% in midday trading Wednesday after Reuters first reported the incorrect response from Bard.
Google unveiled Bard earlier this week in an apparent attempt to compete with the viral success of ChatGPT, which has been used to generate essays, song lyrics, and responses to questions previously searched for on Google. Because of ChatGPT’s meteoric rise in popularity, Google’s management has reportedly declared a “code red” situation for its search product.
Google held an event in its Paris office on Wednesday, detailing plans to use AI technology to drastically change how people search for information online. Google’s announcement came just one day after rival Microsoft unveiled a redesigned version of Bing powered by a more advanced version of the AI used by ChatGPT. (Microsoft is putting billions of dollars into OpenAI, the company behind ChatGPT.)
During the presentation on Wednesday, a Google executive hinted at plans to use this technology to provide more complex and conversational responses to queries, such as providing bullet points listing the best times of year to see various constellations and offering pros and cons for purchasing an electric vehicle.
According to the executive, artificial intelligence technology will pave the way for the “next frontier of our information products.”
