Google’s parent company loses more than $100 billion over wrong AI chatbot answer

Spread the love

Google’s parent company lost more than $100 billion in its market value after chatbot Bard got an answer wrong.

The tech giant’s new AI machine has been released to a team of specialist testers and is yet to be rolled out to the public, but investors wiped a fortune off the value of the search engine’s parent firm Alphabet after experts pointed out Bard’s error while answering a question about space photography.

Promotional material for Bard – Google’s competitor to Microsoft-backed ChatGPT – contained a mistake in the response by the chatbot to the query: “What new discoveries from the James Webb space telescope (JWST) can I tell my nine-year old about?”

Google tweets about its AI chatbot Bard in this screenshot obtained from social media, February 8, 2023. Twitter/Google via REUTERS

Bard’s response, contained in a video demo posted online, included an answer suggesting the JWST was used to take the very first pictures of a planet outside the Earth’s solar system, or exoplanets.

The first picture was actually taken by the European Southern Observatory’s Very Large Telescope in 2004, according to Nasa.

On Wednesday (08.02.23), Alphabet stock opened lower at $102.05 per share against the previous closing price of $107.64 on the Nasdaq.

During the day’s trading session, the tech stock declined as much as 8.9% to hit a low of $98.04, before the share settled 7.68% lower at $99.37 – erasing more than $100 billion in market capitalisation.

The Bard error was picked up by experts including Grant Tremblay, an astrophysicist at the US Center for Astrophysics.

He tweeted about the mistake: “Not to be a – well, actually – jerk, and I’m sure Bard will be impressive, but for the record: JWST did not take ‘the very first image of a planet outside our solar system’.”

Bruce Macintosh, director of University of California Observatories, also tweeted: “Speaking as someone who imaged an exoplanet 14 years before JWST was launched, it feels like you should find a better example?”

Bard is based on a “large language” artificial intelligence model, which is meant to mimic the underlying architecture of the brain in computer form.

It is fed huge amounts of text from the web to teach it how to respond to text-based prompts, but the input can lead to the bot repeating errors from the information it absorbed.

Google said the error, first reported by Reuters, underlined the need for testing new systems, with a spokesperson for the company adding: “This highlights the importance of a rigorous testing process, something that we’re kicking off this week with our trusted tester program.

“We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and groundedness in real-world information.”

Gil Luria, senior software analyst at D.A. Davidson, said “while Google has been a leader in AI innovation” over the last several years, they “seemed to have fallen asleep on implementing this technology into their search product”.