Google co-founder Sergey Brin, has admitted that Google’s Artificial Intelligence (AI) chatbot Gemini “messed up” image generation due to inadequate testing, India Today reports.
Gemini, which was formally named Bard, caused many users irritation when it started refusing to generate images of white people, only generating images of black people and even showing white historical figures as black people.
This sparked a lot of criticism from users, and Gemini soon became a hot talking point, but not in the way Google would have wanted.
Brin, who had stepped away from operations at Google, admitted to coming back because of the allure of AI’s potential. He said this at AGI House in San Francisco.
AGI House is a community of AI founders and researchers committed to helping the general public with the AI transition.
Brin acknowledged that people were right to be upset at Gemini’s messing up of image generation. The co-founder founded Google alongside Larry Page, and was president of its parent company, Alphabet.
He stepped down in 2019 as President, but has remained on the Google board and still owns a relatively large share of the company. He is back in operations at Google to actively contribute to boosting their presence in the AI market.
He recognizes how competitive the AI market is.
Brin added they have suspended the image generation part of Gemini, and are working round the clock to get it to work the way they want it to.
As things stand, you would be unable to generate images, and if you try, you will get a response notifying you that the feature has been temporarily deactivated and you will be notified when it’s in perfect working order.
