Google is reintroducing the ability for users to generate images of people through its Gemini AI chatbot, following a previous suspension due to concerns about historically inaccurate depictions, such as racially diverse Nazis. The company announced that an early access version of this feature will be available to Gemini Advanced, Business, and Enterprise users in English in the coming days.
This capability is powered by Imagen 3, Google's latest AI text-to-image generator. Launched through AI Test Kitchen, Imagen 3 can create everything from photorealistic landscapes to textured oil paintings with just a brief description.
Earlier this year, Google paused the feature after users reported the generation of historically inaccurate images. The updated Imagen 3 model now includes built-in safeguards and has been shown to perform well compared to other image generation tools available, as noted by Dave Citron, Google’s senior director of product management for Gemini.
It's important to note that Gemini will restrict users from creating photorealistic images of public figures, including minors, as well as any gory or sexual content. Citron stated, “Not every image Gemini creates will be perfect, but we’ll continue to listen to feedback from early access users to improve the tool.” Google also plans to expand this AI-generated imagery feature to more users and languages soon.