The Blurred Line of AI in Photography microblog.pratikmhatre.com
@pratik This might be too simple, but as a first step I think separate "Made with AI" and "Edited with AI" labels would help a lot.
@manton Good idea. The "Made with AI" would define a starting point for the source. All AI-generated images (from scratch) must have an invisible digital watermark so browsers and other display apps can detect that and inform the viewer. People who are fine with using such AI images or even viewing them wouldn't mind. Others can decide for themselves what to think of it.
@pratik I can get with an "edited with AI" label depending on its use. I feel like in journalism (which is not my profession) there should be NO AI manipulation whatsoever. In marketing (my profession) I'm comfortable with the lines blurred. I've already used generative AI to do very fast touch-ups when I've been under deadline (huge time saver.) Should THAT get a label? I'd say no. If the image had a large portion using gen AI for the final, I don't know. I haven't come across that issue yet.
@vazquez Agree on all. There are some fields where people are expecting reality. However, journalism has also experienced photography manipulation even before AI, so people who want to deceive will find a way. It's up to those fields and professions to set standards. Even in marketing photography, it's not always (or rarely) the truth. For e.g., the syrup in pancake ads is, in fact, motor oil, and there are pieces of cardboard between the pancakes.