Jessie Clark, our Chief Creative Officer gives her thoughts on AI Photoshop Beta and Design.
AI is here, and it is here to stay. As we learn to navigate this super powerful tool, we are encouraging our team members to spend some time investigating AI as a tool, and explicitly not a crutch. Our Chief Creative Officer, Jessie Clark, gives her take on AI and Design. Here are some of her thoughts.
I began experimenting with AI capabilities in Photoshop Beta. The generative fill option can be a game changer and time saver for designers as we seek to stay abreast of the ever-changing digital landscape. This tool allows the user to provide detailed instructions creating an image per those instructions.
For me, I can instruct Photoshop Beta to create an image I envision in my mind. For example, I instructed the tool to create a “white beach towel on concrete next to a pool with clear blue water.” I then can use the concept as a baseline for design instead of starting from scratch every single time. I also spend copious time searching for stock imagery on Adobe Stock, Freepik, and Shutterstock. I am excited to use this tool to save time searching for imagery whenever I need a generic image.
Photoshop Beta can also be used to change the backgrounds of photos to create different orientations. I can take a photo that was captured in portrait mode and use the “fill in background” tool to create a landscape mode of the same image. This is a fabulous hack for social media, websites, and other digital platforms that require a specific image size.
AI has limitations, and it is truly important to recognize them. Text and human faces are difficult for AI to generate and the technology is not able to fully capture the subtle details and variations in human faces which leads to unnatural and distorted results. Because AI pulls bits and pieces of different imagery together, some faces can end up quite hilarious. (It reminds me of the skit Conan O’Brien did on his late night show, If They Mated, when he would take parts of celebrities’ faces and attempt to predict how their children would look.)
Another limitation of AI is the complexity of language. As a designer, I’ve learned that you must input the language AI understands. I cannot simply tell it to create an image of: “A penguin painting a Picasso” because AI doesn’t decipher between the word “painting” as a noun, and “painting” as a verb. With the request of “A penguin painting a Picasso,” AI generated a Picasso-style painting of a penguin instead of a penguin painting a piece of art. To obtain an accurate image, users must master the language of AI. Continuing with the example, AI better understood: “A penguin in the act of painting a Picasso-style picture” and generated an image I was envisioning.
The AI language requires users to provide a clearer style description in order to get a more accurate picture. Words like cinematic, 3D rendering, painting, poster, minimalist, geometric, futuristic, and photorealistic are all great examples of words to use when instructing AI. These descriptors will help AI generate the style the user is envisioning.
Personally, Ideogram has proven to be a very accurate tool to use to create AI-generated images. This site allows users to create some free concepts and explore what other users are creating. As a designer, this has been a great learning tool for me as I continue to learn and navigate the new world available to me. Specifically, it has helped me learn more about mastering the “AI language” so I can speak the language of AI and get more accurate results.
AI is a tool we are all learning to use. We won’t shy away from it, but we also want you to know that the humans on staff will always be central to creating, writing, and designing marketing content at TealHaus.