Google Search’s generative AI is now able to create images with just a text prompt

Google is taking on Microsoft at its own game as the tech giant has started testing its own image generation tool on the AI-powered Search Genative Experience (SGE).

It works almost exactly like Bing Chat: you enter a prompt directly into Google Search, and after a few seconds four images appear. What’s unique about it is that you can choose one of the images and develop it even further by editing the description to add more details. Google sets the example of asking SGE to generate “a photorealistic image of a capybara” cooking breakfast in the forest. The demo then shows you how to change specific aspects, such as changing the food the animal cooks, from bacon to hash browns, or switching the background from trees to the sky.

This feature isn’t limited to just Google Search, as the company claims you “may see an option to create AI-generated images directly within Google Images.” In that case, one of the image search results will be replaced with a button that provides access to the engine. The creation slides in from the right into its own subwindow.

Limits