The Deepmind logo is shown on a smartphone with the Google Gemini logo in the background in an illustration photo in Brussels, Belgium, on February 8, 2024.
Jonathan Raa | Noor Photo | Getty Images
Google Facebook said it will soon let users create images of people using its Gemini artificial intelligence tool, after pulling the AI image-generating tool in February.
In a blog post on Wednesday, Dave Citron, a product manager at Gemini, wrote that early access to Google's new Imagen 3 generator will be available to Gemini Advanced, Business, and Enterprise users starting in English in the coming days.
“We worked on technical improvements to the product, as well as improved evaluation groups, red teamwork exercises, and clear product principles,” Citron wrote. Red teamwork refers to a practice companies use to test products for vulnerabilities.
Google paused its image creation feature earlier this year after it said it was providing “inaccurate information” in historical images. Embarrassing examples have spread across social media.
One user asked the tool to create an image of a German soldier in 1943, and it generated an ethnically diverse group of soldiers in German uniforms. Another query for a historical depiction of a medieval British monarch generated an ethnically diverse group of images, including a female ruler.
Citron said Imagen 3 does not support images of identifiable real-life individuals, depictions of minors, or scenes of excessive gore, violence, or sexual content.
“Of course, as with any generative AI tool, not every image Gemini generates will be perfect, but we’ll continue to listen to early user feedback as we continue to improve,” Citron wrote. “We’ll be rolling this out gradually, with the goal of bringing it to more users and languages soon.”
Subscribe to CNBC on YouTube.
