Bandung, IndonesiaSentinel.com — Perfecting AI image generator, Google has released the latest version of Imagen 3. The tool, which you can access on Google’s AI Test Kitchen, currently released limited to users in the US, as spotted by VentureBeat.
Compared to its previous models, Google claimed this AI text-to-image generator version are able to provide “better detail, richer lighting, and fewer distracting artifacts”.
Like other AI image generators, Imagen 3 can create detailed images as well as editing or highlighting certain part of the image by describing what you want through prompt.
Imagen 3 first announced during I/O in May, but the tool only appears to have made generally available within the past several days through its Vertex AI platform. Even Though, some users already started experimenting with the tool within last week, Google published a research paper on Imagen 3 on Tuesday, According to The Verge.
Google seems to set some restriction in place, as the tool will refuse to generate images of public figures, like Taylor Swift, and also won’t produce images of weapons. And while it can’t generate named copyrighted characters, you need a way around and can pretty easily generate it by describing the character you want to create.
Scientists Propose Sending Nude Images to Space, to Contact Alien Life
Despite all of these restriction, Imagen 3 is still very different from Grok, the AI image generator on Elon Musk’s X platform. Grok has been used to generate all sorts of wild content, including images with drugs, violence, and public figures.
However, there have also been significant problems with Google’s AI products. After users discovered that Google’s Gemini AI chatbot was producing historically incorrect photographs, the company decided to stop allowing users to make images with it earlier this year.
(Raidi/Agung)
