Photoshop's new generative AI tools explained

Get to grips with the new Generate Image feature and AI tools for jaw-dropping results

- James Paterson

Last year, we saw new generative AI features appear in Photoshop for the first time. These were simultaneously awe-inspiring and oddly underwhelming, with results that could range from the sublime to the ridiculous. Now, inevitably, we’re being treated to new generative AI tools. Powered by Adobe Firefly 3, the big new feature is the Generate Image tool, which like the AI tools we’ve seen from the likes of MidJourney and Dall-E, lets us generate a new image from scratch based on a text prompt.

The challenge for Adobe is how to build generative AI into Photoshop tools without the AI just taking over. As photographers, we don’t necessarily want a button that lets us make a generically good image. But we might want to augment and build upon an image that we’ve shot. The new features in Photoshop Beta go some way to addressing this. The Generative Fill and Expand tools now allow us to influence the AI-generated content by including a reference image, so we can upload a photo of our own to guide the tool.

These ground-breaking features continue to push the boundaries of image-making. However, you will need lots of patience – it usually takes several aborted attempts and a slew of dodgy fills before something good happens. Here are some of the key features of the new AI tools…