Google search has learned to generate images

Google continues experiment with the introduction of generative artificial intelligence into their services. Now the turn has come to the search, in which the neural network will be able to create images based on a regular request.

For example, you want to make a card for a friend who loves to cook and whose favorite animal is a capybara. Just search for “draw a picture of a capybara wearing a chef’s hat preparing breakfast.”

The neural network will create several images to choose from and show a prompt request, which, if necessary, can be edited for a new generation.

You can also use generative AI when searching for images. Google will offer this feature right in the results. Below is an example of a “minimalist Halloween table setting.”

AI, of course, does not work with harmful or misleading content. Each generated image also receives metadata tagging and embedded watermarks indicating that it was generated by a neural network.

The image generation feature is currently only available in English in the US as part of testing. Later, the function may begin to be implemented worldwide.