Testing out the new ‘Generative fill’ in the Photoshop Beta I have got to say is pretty impressive. 

I took some Photoshop landscape illustrations I created that are a few years old. Which were created as vertical format illustrations. This one was commissioned for a T-Shirt design. I mirrored the artwork over to the right and then used Generative Fill to change the layers. This took quite a few minutes for each one. As you can see it works incredibly well, a big time saver.  At the moment I am only able to get it working for specific types of illustrations such as these landscapes. Other images it seems to want to style change, which doesn’t fit with the rest of the illustration.

My anatomical illustrations for example could not change eyes around for some reason. Plus figurative silhouettes kept coming out a mess. I guess in the future there will be different solutions to the varying types of image-making styles. It surprises me that AI has not been trained on anatomy very well it seems.

Being able to add small extra elements via a simple text prompt such as splash, mountains, and trees, and add red mist is a significant addition to ‘photomontage’ if we can call it that. It occurs to me that PS is in danger of becoming something completely different to what it currently is. The convergence of what is hand crafted and automated is a fascinating conflicting and constantly changing debate. The rug has truly been pulled from underneath many of us.

I still don’t use it for my professional commissioned work , I love taking time and having creative control over every aspect, being able to put my mark on every pixel or drop of ink give me so much more pleasure and reward. The best creative results come from taking time to get to know the subject matter, the people and company I am working for.  

Let me know what you think?