The Bard chatbot, created by the Bard Corporation, attributed a discovery to the James Webb Space Telescope that it did not make. During this time, the market value of the company on the stock exchanges decreased by USD 100 billion.
Google’s chatbot made a mistake
Google lost almost USD 100 billion on global exchanges in one day. The reason for this is just one post on Twitter, which demonstrates to everyone how the artificial intelligence created by this corporation is mistaken in a not too complicated issue related to space.
We are talking about the Bard chatbot, which Google has developed and is now actively promoting. The question it couldn’t handle perfectly looked like this: “What new discoveries of the James Webb Space Telescope can I tell a 9-year-old child about”
The Bard’s answer was partly correct. It is true that it named the “green pea” galaxies and star systems 13 billion years old as an example. But then it was attributed to James Webb obtaining the first ever images of exoplanets, and this was incorrect.
What was the mistake of artificial intelligence?
There were indeed reports that James Webb had discovered exoplanets. But these were the first objects discovered by it. In general, the first image of the disk of a substellar object was obtained almost 20 years ago. In 2004, it was made by the Very Large Telescope of the European Southern Observatory in Chile.
This exoplanet is called 2M1207b. It orbits a brown dwarf, which is an object larger than a planet, but not big enough to support nuclear fusion and shine like a star. The image was a great success: the reflected light of the planet was taken from a distance of 230 light years.
Competitors are also wrong
According to Reuters, it was because of this small mistake that the parent company of the search giant, Alphabet Inc., lost USD 100 billion in market value on Wednesday. Industry observers worry that Google is rapidly losing trust amid the success of rival bot developer OpenAI, a startup that actively supports the search giant Microsoft.
It was they who, back in November 2022, released the ChatGPT artificial intelligence, which can create very natural texts and even writes in different styles. However, it also regularly makes mistakes concerning simple scientific issues.
For example, last week a bot made a mistake related to a fundamental equation that was used in rocket science. ChatGPT was asked to name it, and it gave an answer, but a specialist in this field who was present in the studio showed that it was not correct.
According to www.space.com
Follow us on Twitter to get the most interesting space news in time