Google is set to reintroduce the capability to generate human images in its Gemini AI model, which previously faced criticism for producing historically inaccurate human depictions. Earlier this year, the Gemini AI model sparked controversy when it created images of people that, while realistic, were historically inconsistent and appeared to have had commands covertly altered.
This resulted in odd images such as Chinese Vikings and Black Nazis. In response to the backlash, Google temporarily restricted Gemini’s ability to generate human images. However, Google is now preparing to re-enable this feature for Business and Enterprise users of Gemini Advanced, with support available in English.
With Imagen 3, Gemini will be able to create human images again
Google expects the issues with Gemini to be resolved with this upgrade, powered by Imagen 3. Imagen 3 was quietly launched earlier this month via AI Test Kitchen. This advanced tool can create anything from photo-realistic landscapes to watercolor paintings with just a few words. Google suspended Gemini’s ability to generate images of people in February after discovering that users were creating historically inaccurate depictions. The updated Imagen 3 model, however, will include built-in security measures. According to Google, it delivers more accurate results than existing image models.
Gemini will not permit users to create photo-realistic images of public figures, content involving children, or scenes that are gory, violent, or sexual in nature. Google plans to gradually make this capability available to more users over time.,
You may also like this content
- Lionsgate and Runway: Revolutionizing Film with AI
- Amazon Launches AI Assistant to Support Sellers Effectively
- Microsoft and Google Explore Mini Nuclear Reactors for AI
-
-
-
-
-
-
-
-
-
-
