Google has admitted to inaccuracies in the historical image generation feature of its AI chatbot Gemini.
Following recent issues, the company has temporarily stopped Gemini from generating images of people and is actively working to rectify the situation.
This move comes after users shared screenshots revealing racially diverse characters appearing in historically white-dominated scenes generated by Gemini.
Critics have raised concerns about whether Google is over-correcting for racial bias in its AI model.
In response, Google stated that they are addressing the recent issues with Gemini’s image generation feature and have paused image generation of people while they work on improvements.
Previous studies have highlighted how AI image-generators can amplify racial and gender stereotypes present in their training data.
While Gemini can generate a wide range of people, Google acknowledges that it is currently missing the mark.
The company expects to release an improved version of the image generation feature soon.
Gemini’s response to generating pictures of people indicated that it is actively working to enhance this ability, and users can expect the fe
ature to return soon.
