This photo illustration was taken in Brussels, Belgium on February 8, 2024, and shows the Deepmind logo displayed on a smartphone with the Google Gemini logo in the background.
Jonathan Lahr | Nurphoto | Getty Images
Google, which stopped offering its AI image generation tool in February, has announced that it will soon allow users to create images of people using its artificial intelligence tool Gemini.
Early access to Google's new Imagen 3 generator will be available to Gemini Advanced, Business and Enterprise users in the English language within the next few days, Dave Citron, senior product director at Gemini, wrote in a blog post on Wednesday.
“We have been working on technical improvements to our products, improving our assessment set, red teaming, and developing clear product doctrine,” Citron wrote. Red teaming refers to a methodology companies use to test their products for vulnerabilities.
Earlier this year, Google suspended its image generation feature due to “inaccuracies” in historic photos after embarrassing examples were spread on social media.
One user asked the tool to generate images of German soldiers in 1943, and it produced a racially diverse set of soldiers in German military uniforms. Another query, asking for historical depictions of medieval English kings, also produced a racially diverse set of images that included female rulers.
Citron said Imagen 3 does not support realistic character identification, depictions of minors or excessively gory, violent or sexual scenes.
“Of course, like any generative AI tool, not all images Gemini creates are perfect, but we will continue to improve it as we listen to feedback from our early users,” Citron wrote. “We aim to gradually roll this out and make it available to more users and languages soon.”
Subscribe to CNBC on YouTube.
