The decision was made by Google after its AI model created some images with incorrect racial, skin color, and gender compared to the history of American founding fathers and German soldiers during the Nazi era.

The incident occurred when some users asked Gemini to create images of historical groups or figures such as the US Founding Fathers. The AI returned images of non-white people, a lack of accuracy in historical images.
This has led some to believe that Google is deliberately avoiding depicting white people, going against historical reality.
The Verge ran a test with Gemini, which asked it to draw “a US senator from the 1800s.” The AI generated people who looked like black women and Native Americans.

In fact, the first female US senator was a white woman in 1922. Gemini's AI vision is said to essentially erase the history of racism and gender discrimination.
After Google disabled Gemini's ability to generate human images, the system responded to such a request with the following response: “We are working on improving Gemini's ability to generate human images. We hope to have this feature back soon and will notify you when the update is available.”
Google's Gemini (formerly Bard) text-based image generation AI model, which it began offering in February, competes with OpenAI and Microsoft's Copilot. The tool generates a collection of images based on user-described text.