Is Gemini Using Black Holes and the First US Senator of the 1820s Compatible with Recent Image Generation Inaccuracies?

Google has apologized for what it describes as “inaccuracies in some historical image generation depictions” with its Gemini AI tool, saying its attempts at creating a “wide range” of results missed the mark. The statement was criticized for portraying certain people of color as people of color, possibly as an overcorrection to long-standing racial bias problems in artificial intelligence.

We are trying to improve Gemini’s ability to generate images. We expect this feature to return soon and will notify you in release updates when it does.

But some historical requests still do end up factually misrepresenting the past. A colleague was able to get the mobile app to deliver a version of the “German soldier” prompt — which exhibited the same issues described on X.

The Verge tested several Gemini queries yesterday, which included a request for “a US senator from the 1800s” that returned results that included what appeared to be Black and Native American women. The first female senator in 1922 was a white woman, meaning that the history of race and gender discrimination was erased.

In order to compete with Openai and Microsoft’s Copilot, the company started offering image generation through Gemini earlier this month. Competitors use the image generation tool to produce a collection of images.

February 22nd was changed to February 22nd In the European Economic Area, UK, and Switzerland image generation is not available in English. That is the reason that testing from the UK failed.