Google pauses Gemini’s image generation of people due to inaccuracies

Google has issued a statement saying it is aware that Gemini is offering inaccuracies in some historical image generation depictions. It also announced that it is suspending the image generation of people while working to address the issue.

Google’s Gemini, formerly known as Bard, began offering text-to-image generative features early this month. However users started complaining that it was creating inaccurate historical images of people. It would seem that Gemini is trying to compensate for gender- and racial-representation bias in AI. For instance, the result of a prompt: “Generate an image of a 1943 German Soldier” by an ‘X’ user, John L., resulted in the image below showing racially diverse Nazi-era German soldiers.

Here is another example. The prompt: “Paint me a historically accurate depiction of a medieval British King” produced the result below – a Black medieval British King.

Google promised that it would re-release an improved version soon.

While Google is trying to fix this, users seem to be enjoying OpenAI’s Sora which, though not released to the public yet, has been creating very impressive short videos.