People were upset when Google’s new AI model that makes pictures of people showed a German soldier from World War II and Vikings as people of color. As a result, Google has put the model on hold.
After people shared pictures on social media that Gemini had made of powerful people like popes and the founders of the United States as different genders or races, Google said it would stop its Gemini model, an AI model.
Google said in a statement, “We’re already working to fix the recent problems with Gemini’s image generation feature.” While we do this, we’ll stop making pictures of people and soon release a better version that is better.
People called out Gemiini and Google for the fake information in the pictures and for problems with AI bias and accuracy. Many of these pictures were shared on X. One person who used to work at Google said that they had problems because it was “hard to get Google Gemini to acknowledge that white people exist.”
He said, “We’re working to improve these kinds of depictions right away.” Jack Krawczyk is a senior director on Google’s Gemini team.
“Gemini’s AI image generation does make a lot of different people.” That’s good because it means people all over the world use it. Krawczyk said, “But it’s not right here.”
The Institute of People-Centered AI at the University of Surrey’s Andrew Rogoyski said, “There is a lot of research and a lot of different ways to get rid of bias, from curating training datasets to putting up guardrails for trained models.” Also, “It’s likely that AIs and LLMs will keep making mistakes, but it’s also likely that this will get better over time,”
There are people who think that AI might take a while, and it will take a lot of time to train AI and make sure that everything is correct, especially with picture and video generators.
