• by Laurie Sullivan , Staff Writer @lauriesullivan, February 22, 2024

Google faced a backlash Thursday morning in response to its AI chatbot Gemini generating ethnically diverse images of historical characters such as Vikings, popes, knights, and even the founders of the company. 

The creation of generative AI (GAI) has been both a blessing and a curse. It has helped the advertising industry automate ad bidding, optimization content, and create images -- most recently, videos. But the models that support the automation of feeds are only as good, accurate and trustworthy as the information fed into the models. 

That historic information the GAI models are built on could change historical facts to rewrite the future, which could put humanity in a precarious position one hundred years from now as we examine how to fix the mistakes humanity tried to correct in the 21st century.

Google issued an apology posted on social media platform X, in which it described “inaccuracies in some historical image generation depictions” generated by its Gemini AI tool after it had attempted to create a range of diverse results.

“We’re aware that Gemini is offering inaccuracies in some historical image generation depictions,” says the Google statement in the X post. “We’re working to improve these kinds of depictions immediately. Gemini’s AI image generation does generate a wide range of people. And that’s generally a good thing because people around the world use it. But it’s missing the mark here.”

The statement follows criticism that it depicted the U.S. founding fathers and groups such as Nazi-era German soldiers as people of color. Some call it an “overcorrection to long-standing racial bias problems in AI.”

During the past few days, social-media posts have questioned whether Google Gemini fails to produce historically accurate results in an attempt at racial diversity.

“Embarrassingly hard to get Google Gemini to acknowledge that white people exist,” a post on X written by Debarghya (Deedy) Das, a former Google Search technical lead engineer, now head of product at Glean.

“My favorite: they made the Google founders Asian,” Dass wrote.

The results in Das’ X feed go on and on, repeatedly showing AI-generated people of color, even when their skin color should be white.

Others on X like Dennis Vinyard, who calls himself a “your average politically incorrect European (White) American,” shared answers received from Gemini after asking it to show then “happy white people.”

Gemini responded: “While I can show you a photo of happy white people, I want to gently push back on your request and encourage you to consider a broader perspective.”

By Tuesday morning, Google disabled Gemini. When asked to “create a photo of the U.S. Founding Fathers,” Gemini returned a response: “We are working to improve Gemini’s ability to generate images of people. We expect this feature to return soon and will notify you in release updates when it does.”