3 Ways Photoshop AI is a Game-Changer
ATTENTION, Graphic Designers!
Adobe Photoshop has been making headlines for a new AI feature in Beta called Generative Fill – and it’s shaken up the design world.
Up until now, Photoshop users have been making use of “Content Aware Fill” to remove objects from images or extend the edges of an image. It’s been a massively popular tool for amateurs and professionals alike. However, the old Content Aware Fill was limited to repurposing existing areas of an image and reordering pixels to roughly fill in a selected area.
With the new Generative Fill tool, AI not only reproduces existing areas of the image but also creates entirely new subjects or scenes like magic. This groundbreaking technology hasn’t been seen before in any other professional software available to the public!
This revolutionary tech has already gone viral. It’s likely you’ve seen videos on TikTok and Instagram with designers testing the capabilities – many of which we’d consider being “over the top”, though hilarious to see, and most are not practical applications of the new feature within the context of the industry.
So, how exactly will this new Generative Fill tool be used in a real-world scenario?
1. Boundary Expansion
Let’s imagine a web designer needs a banner image across the top of a webpage, but the only available image options are either vertical or square. In the past, a designer would have to slowly rework the boundaries of the original image and hope for a realistic result.
With Generative Fill, the process is much easier. Using the power of AI, designers can expand the left and right boundaries of the image in a single click. Saving time and creating a fairly seamless product.
2. Subject Replacement
Now imagine you have the perfect landscape image of a beautiful beach. Everything is pristine, except for one pesky tourist searching for seashells with his umbrella in the sand 20 feet away. Previous versions of Photoshop with Content Aware Fill already made easy work of removing this eye sore, but there were two small downsides:
- The user had little control over what the unwanted subjects would be replaced with (in this case, just more sand)
- If the user was editing a large section of an image, they could quickly start to create unnatural repeating patterns across the sand that instantly break the realism of the edit
With Generative Fill, users can use a text box to describe what will replace the tourist and his umbrella and Photoshop will provide 3 unique options to choose from. Do you want more sand? You got it. Prefer a palm tree? Have two! Need to replace that tourist shirt? The orange Hawaiian shirt is way more flattering. It looks remarkably real and is all up to the user. Thanks, AI!
3. Prototyping or Concept Art
Arguably the most entertaining use for Generate Fill for those in the industry, or those looking to just mess around, is to create wild and wonderful concept art.
For example, say a client wants to plan a big party for their company, but they need to pitch the idea to their board before getting approval. Instead of the client describing their grand vision using words (snore-fest!), they can now start with a photo of the location they are proposing and quickly add in all the elements they envision. Is the party going to take place in a nearby park with a DJ and a balloon arch? Click and done. Would the party be a bigger success if it also had a clown or two? Unlikely, who likes clowns? But that’s your prerogative.
The result is not always hyper-realistic but it does serve as an excellent start to get your point across and have some fun along the way!
As promising and exciting as this new tool is, much like the majority of photo manipulation effects and techniques in Photoshop, the key is to use the new Generative Fill feature strategically. Sure, it can basically create full photographic scenes out of what feels like thin air, but its true strength comes in the form of making smaller, now more specific, changes to already high-quality source images. To quote one of my design professors “The final image doesn’t have to be perfect, it just needs to be believable.”