- Google’s Search Generative Experience can now create images from text.
- The feature is available now, but it may take some time to reach everyone.
- Google says that it’s putting safeguards to prevent misuse of the feature.
- You can access the feature through SGE, not Bard.
Google is updating its Search Generative Experience (SGE) to add the ability to create images using AI. The search giant has announced that it’s bringing the image creation feature slowly to users who opted into the Search Lab program starting October 12, 2023.
According to the company, to create an image with AI, you will only need to type a descriptive query of the image you want. As you submit the query, SGE will analyze the request and create four variants of the images based on the text.
You can choose one of the images to expand the view and access the “Edit” option to modify the images further using prompts to the Search Generative Experience.
You may also see the option to create AI images directly from the Google Images page while searching for inspiration. If the feature is available, you will notice the “Create something new” with the “Generate” button as you scroll through the results to create a new image with AI.
It’s important to note that this feature is available through the search page experience, not from Bard AI, and it uses the Google Imagen text-to-image model capable of “photorealism and deep language representations” instead of the Adobe Firefly integration.
I’m making this clarification as, in the past, Google has said it was partnering with Adobe to bring its Firefly technology as an extension for Bard to create high-quality images that you can edit with Adobe Express. However, this is a different project.
The announcement also points out that Google is making the best effort to roll out this feature with different safeguards to prevent the creation of images that are not permitted per the company’s policies. Furthermore, every image created using the tool includes metadata and watermarking to make it clear that the image was created using AI.
Although the ability to create images with AI is new to the Google Generative Experience, Bing Chat has included a similar feature for some time through the integration of the DALL-E language model from OpenAI. Even further, Microsoft has already brought the feature to Windows 11 through Copilot and with the Cocreator feature on Microsoft Paint.
Google also revealed that it is gearing up to release a new tool known as “About this image” to help users get more content and credibility of images.
When using SGE to create images, it’s essential to make clear the intent by including action keywords like “draw a picture” or “create an image” in the query and provide as much description as possible for a more accurate result.
The ability to create images using AI is available now, but it may take some time to reach everyone, and remember that you must opt into the Search Lab to access the feature.