The use of generative AI technologies, including Photoshop's Generative Fill , ChatGPT, and Stable Diffusion are rapidly becoming more ubiquitous among creatives. My thoughts about using AI are rapidly changing as the technology evolves.
In high school, I worked as a photo editor and Web developer at a small photo studio in Middleton, MA . At one point, I was given the task to remove a family from a background of a wedding photo, leaving just the bride and groom. That edit took me several careful hours with Photoshop CS2's clone stamp and adjustment tools. After I finished editing that photo - a significant portion of which was digitally altered - it was printed on the front page of a local wedding magazine. I was very proud.
With access to Photoshop's Generative Fill, teenage Zach would have finished that project in seconds, not hours, and the result would have been truly impossible to discern from a real photo. Given the myriad ethical conundrums with using AI models whose source images aren't always used with permission, would that have been OK?
Today, while editing portfolio or assignment photos, I rarely use Generative Fill. Usually, I use it to remove distracting background objects. Every time I do so, I wonder: Does my use of this AI tool cheapen or diminish the impact of my final output, even if nobody will recognize my usage of the tool? These thoughts closely resemble discussions about the publishing of composite photos .
At some point, I'm going to have to figure out how to ask clients if they're comfortable with any use of generative AI in their projects. In order to do so, I'll have to figure out a quick way to help folks understand the different sides of the story.
Do you have any thoughts about the use of generative AI in photography? You may be interested in one of my AI-related projects, I Am A Good Bing 😊 , a news website I co-developed with my wife that aggregates silly and serious stories about modern AI. If you'd like to chat more about this, contact me.