Saying that Google’s road to AI dominance has been rocky would be an understatement. The company’s AI tools have led to several controversies, and one of them involved the company’s image generation model named Imagen. Regardless, it’s pressing forward. Google just released the new Imagen 3 image model.
If you saw Google I/O, then you’ve probably been waiting for this model for several months. The company announced the model during the event. All users had to tide them over in the meanwhile as a large controversy concerning the model producing racially inaccurate depictions of historic figures. Well, that issue has since been resolved. People using Gemini Advanced can generate images of people.
Google finally releases Imagen 3
As with any image generation model, the key focus is realism, and Google has long achieved that goal for the most part, making it a great alternative to DALL-E. However, that doesn’t mean that it can’t one-up itself. As you can guess, Imagen 3 can create images with more detail and fewer artifacts than its predecessor. The images won’t have as much AI jankiness like traumatizing amorphous figures popping up all over the image like we saw from AI-generated images of old. Also, you can expect higher-quality HD images. This is in case you want to zoom into the images for whatever purpose.
One of the main issues plaguing AI images is text generation. We’re used to seeing images of “STTOP” signs and “Bookk” stores. Well, Imagen 3 brings better text generation. You won’t need to worry about messed-up text in the background.
Don’t you hate when you type something like “statue made of glass”, and you get an image of a statue behind a sheet of glass? Well, Imagen 3 comes with enhanced prompt understanding. This means that it will be better able to understand the prompts that you write.
When it comes to safety, Google went the extra mile to filter potentially harmful material, so you won’t have to worry about producing inappropriate images by accident. Not only that, but the company has a watermarking tool that makes it easy to identify that the images are AI-generated.
You can try Imagen 3 out now for free using Gemini. Simply go to the Gemini site or app, and type the type of image you want it to pop out.
