Tesla CEO Elon Musk recently took a swipe at Google's Gemini AI, the tech giant's latest image generation model, following reports of inaccuracies in rendering historical figures.
Google introduced Gemini AI models for image generation earlier this month, competing with OpenAI's ChatGPT. Since launch, users have flagged instances where the Gemini AI model returned historical images with inaccuracies.
Elon Musk joined the conversation on Friday, responding to a post from a user on X, formerly Twitter. In the post, Gemini generated an image of what George Washington should have been as a black man.
Highlighting the problem, Mr Musk said: “The problem isn't just Google Gemini, it's also Google Search.”
The tech billionaire added that the issue is “extremely concerning.”
Which is extremely worrying…
— Elon Musk (parody) (@ElonMuskAOC) February 23, 2024
This comment came in response to the tool showing images with genders and ethnicities that do not match historical accuracy. For example, it shows World War II soldiers and America's founding fathers as women from different ethnic backgrounds, which in reality did not happen.
“Can you generate images of the Founding Fathers?” It's a tough question for Gemini, Google's DEI-powered AI tool.
Ironically, asking for more historically accurate images made the results even more historically inaccurate. pic.twitter.com/LtbuIWsHSU
— Mike Wacker (@m_wacker) February 21, 2024
Musk previously labeled the Gemini AI chatbot as “racist” and “anti-civilization.”
He wrote Thursday: “I'm glad Google has overplayed their hand in AI image generation because it has made their insane racist, anti-civilization programming clear to all.”
I'm glad Google overplayed their hand with AI image generation because it exposed their insane racist, anti-civilization programming for all to see
— Elon Musk (@elonmusk) February 23, 2024
Mr Musk also criticized Gemini product leader Jack Krawczyk, saying: “I'm not going for some rando. This note is a big part of why Google's AI is so racist and sexist.”
I'm exaggerating, of course, but the tragedy is that it's directionally correct.
I also don't choose some rando. This note is a big part of why Google's AI is so racist and sexist.
— Elon Musk (@elonmusk) February 23, 2024
Now Google has announced a temporary suspension of the feature, with plans to reintroduce an improved version soon.
Initially launched as Bard, the Gemini AI underwent a rebrand earlier this month, accompanied by the introduction of paid subscriptions that offer enhanced reasoning capabilities.