Google Pauses AI-Made Images of People After Race Inaccuracies

The Gemini logo on a smartphone.

(Bloomberg) — Alphabet Inc.’s Google said it will pause the image generation of people for Gemini, a powerful artificial intelligence model, after criticism about how it was handling race.

“We’re already working to address recent issues with Gemini’s image generation feature,” Google said in a post on X on Thursday. “We’re going to pause the image generation of people and will re-release an improved version soon.”

The misstep comes as Google has increasingly focused on AI after progress by rivals Microsoft Corp. and OpenAI potentially threatened its core internet search business. Advances in AI have led to concerns that the technology will enable deepfakes, misinformation and bias.

Earlier this week, Google said it was aware Gemini was offering what it called “inaccuracies in some historical image generation depictions.”

The move came after critics on X posted numerous examples from the model that appeared to feature historically inaccurate images in regards to the race of the subjects.

The updated AI model, called Gemini 1.5 Pro, was released last week to cloud customers and developers so they can test its new features and eventually create new commercial applications. Google and its rivals have spent billions of dollars to ramp up their capabilities in generative AI and are keen to attract corporate clients to show their investments are paying off.

(Updates with details starting in first paragraph.)

Most Read from Bloomberg

©2024 Bloomberg L.P.