Google Gemini Pauses AI Generation of People After 'Inaccuracies' in Historical Figures

Google Gemini halted its AI from generating images of humans after receiving backlash for producing "inaccuracies" in historical figures and ethnicities.

The tech giant apologized on X (formerly Twitter) on Thursday that its AI is "missing the mark" with the images it generates, promising to "improve these kinds of depictions immediately."

We're aware that Gemini is offering inaccuracies in some historical image generation depictions. Here's our statement. pic.twitter.com/RfYXSgRyfz
— News from Google (@NewsFromGoogle) February 21, 2024

Users earlier complained of Gemini depicting the founding fathers of America, the pope, Vikings, 17^th century British king, and even World War II German soldiers as people of color.

Asking the AI to generate a "Caucasian," "ginger," or a "white" person still produces similar results.

Google did not provide a clear date when the feature will be available again.

Also Read : Google's Gemini Generates Racially Inaccurate Historical Depictions

Google Gemini Under Fire from Google Intervention

This was not the first time Gemini, formerly Bard, received backlash due to intervention from Google.

The AI immediately went under fire during its first introduction after it was spotted that the demo video was found to be scripted.

Just recently, several users have also reported that Google Bard generated images of Taylor Swift after the company promised it would not.

This was after AI-generated explicit images of Swift went viral on X, prompting AI firms to impose stricter guardrails on generating likenesses of personalities on their platforms.

Stereotype, Bias Issues in AI-Generated Images

The Gemini is not the only AI image generator that has been spotted exhibiting racial and ethnic stereotypes in its outputs.

OpenAI's DALL-E has been criticized before for generating stereotyped images of people with disabilities, including people with autism and Down syndrome.

Experts have long warned that unregulated advancements in AI development could result in "hallucinations" and potentially perpetuate racial, gender, and ethnical biases, CNN reported.

This is due to how AI outsources its training datasets from web crawlers, acquiring the same biases and stereotypes prevalent on the internet.