In the early days of last December, we had to go to special websites to turn our natural language commands into artificial intelligence art. But there is no need for this anymore. In its statement, Google announced that users who prefer Search Generative Experience (SGE) will be able to create artificial intelligence images directly from the standard Search bar.
Rather than selecting websites from a returned list, the system synthesizes a coherent response to the user’s natural language prompt using the same data driven by the list’s links. The system returns rendered images instead of rendered text, a natural extension of this experience. When users type in a description of what they’re looking for (in Google’s case, a Capybara cooking breakfast), the engine generates four alternatives within minutes to choose from and refine further. Users can also transfer or download the images they create to Drive.
What’s more, users will be able to create images directly in Google Images. So if you’re looking for “minimalist halloween table setting” or “spooky dog house ideas” you’ll be able to type that into the search bar and have Google create an image based on it. What’s really cool is that you can run Google Lens on this generated image and search for real products that most closely resemble the image the computer created for you.
Of course, there are some limitations imposed by new features. To use any of these, you must be registered with Google Labs and have chosen to participate in the SGE program. Additionally, the new rendering functions are currently only available within the US, on English-language apps, and for users over 18.
The company is also increasing efforts to prevent misuse of its generative AI technology. Users will be prevented from creating photorealistic images of human faces. If you want a photorealistic Capybara cooking bacon, that’s no problem. If you want a photorealistic Colonel Sanders cooking bacon, you’re going to run into problems. You will also be prevented from creating images of important people.
Finally, Google is also implementing the SynthID system developed by DeepMind and announced last month. SythID is a visually undetectable metadata watermark that not only identifies a created image as such but also provides information about who or what created the image and when. The new features will be rolling out to select users today through the Labs tab in the Google app and Chrome desktop on iOS and Android, and will roll out to more users in the coming weeks.
4014
