Google has decided to discontinue Gemini’s people imaging tool, its generative artificial intelligence. The reason is the historically inaccurate errors of some illustrations and the subsequent criticism of users.
For example, a prompt of “German soldiers in World War II” resulted in artwork featuring Asian female soldiers. It also featured non-white people in a command related to the so-called Founding Fathers of the United States, that is, the main political leaders of the Declaration of Independence. They were all male and white.
By discontinuing the tool that creates images of people in Gemini, Google also said that it is working to resolve the situation. The big tech commented that it expects the service to be back soon.
Anyone who wants to generate images of other subjects, such as landscapes, scenery, products, or to return to the fashion of the Pixar versions, can continue using Gemini without any problems.
Gemini’s imager was released in February 2024, shortly before the company adopted the new name and shut down Bard for good. It’s Google’s answer to the services provided by OpenAI and Microsoft.
The former is the creator of Dall-E, while the big tech founded by Bill Gates used the partnership with OpenAI to bring AI image creation to Copilot (formerly Bing Chat).
Thursday’s decision is a setback in Google’s race to gain ground in the segment.