Google is searching for ways to reassure people that it is still out in front in the race for the best artificial intelligence technology.
And so far, the internet giant seems to be coming up with the wrong answer.
An advert designed to show off its new AI-bot, showed it answering a query incorrectly.
Shares in parent company Alphabet sank more than 7% on Wednesday, knocking $100bn (£82bn) off the firm’s market value.
In the promotion for the bot, known as Bard, which was released on Twitter on Monday, the bot was asked about what to tell a nine-year-old about discoveries from the James Webb Space Telescope.
It offered the response that the telescope was the first to take pictures of a planet outside the earth’s solar system, when in fact that milestone was claimed by the European Very Large Telescope in 2004 – a mistake quickly noted by astronomers on Twitter.
“Why didn’t you factcheck this example before sharing it?” Chris Harrison, a fellow at Newcastle University, replied to the tweet.
Investors were also underwhelmed by a presentation the company gave about its plans to deploy artificial intelligence in its products.
- Google launches ChatGPT rival called Bard
- Microsoft unveils new Bing with ChatGPT powers
- ‘Google killer’ ChatGPT sparks AI chatbot race
Google has been under pressure since late last year, when Microsoft-backed OpenAI unveiled new ChatGPT software. It quickly became a viral hit for its facility in passing business school exams, composing song lyrics and answering other questions.
Microsoft this week said a new version of its Bing search engine, which has lagged Google for years, would use the ChatGPT technology in an even more advanced form.
Though investors have embraced the push for artificial intelligence, sceptics have warned rushing out the technology raises risks of errors or otherwise skewed results, as well as issues of plagiarism.
A Google spokesperson said the error highlighted “the importance of a rigorous testing process, something that we’re kicking off this week with our Trusted Tester programme”.
“We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and roundedness in real-world information,” they said.