How To Use Google's New AI Image Generator in Search (2024)

Never one to be left in the dust, Google has just begun testing an AI generation tool that can create images using a text prompt. Here’s how to use Google’s new AI image generation tool.

Set to rival Microsoft’s Bing Chat, the AI-powered Search Generative Experience (SGE) is currently only available to a small number of American users. Testers must have opted into the SGE program via Google Labs and be at least 18 years old, although that’s still no guarantee you’ll be amongst the lucky few able to experiment right now.

Still, the world of AI image generation has just opened up even more beyond tools like OpenAI’s DALL-E 3. Fortunately, it’s all pretty straight forward. Here’s everything you need to know about creating AI images direct in Google.

How To Get Started with Google AI Image Generation

If you’re one of the lucky few who’s got access to Google AI image generation, then getting started couldn’t be easier. Provided you’re already signed up for Google’s SGE testing program, as mentioned above, all you need to do is:

1. Open up a Google search and enter an image generation prompt

2. Wait a few seconds

3. Take a look at the four image options shared by the SGE

It’s that simple. If you want to edit the images further, select one of them then amend the description to add more detail.

Google has given the surreal example of a “photorealistic image of a capybara cooking breakfast in the forest” to illustrate how it will work. The example then goes on to show an edited description that changes specific details, like asking for the breakfast to become hash browns instead of bacon, or swapping the background trees to sky. Anyone familiar with entering ChatGPT prompts will be used to the idea, but even if you’re not, you should find the process easy enough.

Are There Any Restrictions on Google AI Image Generation?

Well, it wouldn’t be AI generation without a few restrictions.

The SGE features safeguarding measures that block banned content as outlined in the company’s generative AI policy. This includes anything that creates misinformation, promotes illegal activities, and generates sexually explicit content (if it’s not labelled educational or artistic).

You’ll also find that every AI-generated image will come out with a watermark and ‘metadata labelling’ tag. This is to show everyone that content was made by AI, hopefully deterring the spread of false information.

Along those same preventative lines, Google plans to give all its AI-generated content an ‘About This Image’ description. This will aim to supply context about what a user is looking at, so hopefully nobody can claim the AI-images or text are real.

What Else Should I Know About Google AI Image Generation?

Although not yet confirmed, it’s likely that Google’s AI-image generation will become available outside of Search. As in, you’ll see an option to create an AI picture within Google Images too. This could mean one of the image search results is replaced with a button leading to the generative functionality, which will then pop up in a sub-window.

As mentioned above, the feature is currently only available to American users in English. Nothing about an international release has yet been confirmed.

Keen to try and give it a go? You’ll have to sign up to the Search Labs waitlist first. Head over to Google Labs to get going.