Tech giant says model 'missed the mark' after controversy over failing to portray white people
Google has temporarily stopped generating images of people after its Gemini AI model failed to portray white people.
The search engine giant made the announcement after Gemini users on Thursday shared images created by a sample of mostly people of color, including scenes from history that featured only white people.
“Gemini's AI image generation creates a diverse population. That's usually a good thing because people around the world use it. But it missed the mark here,” Google said in a post on X.
“We are already working to resolve recent issues with Gemini's image generation feature,” the tech giant added. “While we're doing this, we're going to pause people's image creation and release an updated version as soon as possible.”
It's embarrassingly difficult for Google Gemini white people to admit they exist pic.twitter.com/4lkhD7p5nR
— Dee Dee (@debarghya_das) February 20, 2024
Gemini-generated images circulated on social media in recent days sparked widespread ridicule and outrage, with some users accusing Google of being “woke” to the detriment of truth or accuracy.
Among the films attracting criticism were the depiction of four Swedish women, none of whom were white, and footage of black and Asian Nazi soldiers.
“It's embarrassing that Google Gemini acknowledges the existence of white people,” Deborkia Das, a company engineer at enterprise search startup Klein, said in a post on X, along with several images generated by Gemini.
AI models have faced criticism in the past for overlooking people of color and perpetuating stereotypes in their results.
Google, which has been racing to catch up with rival OpenAI since the launch of ChatGPT in 2022, has faced several setbacks in releasing AI products.
Last year, the tech giant apologized after its AI chatbot Bart mistakenly claimed during a demo that the James Webb Space Telescope had taken the first images of a planet outside the solar system.
