On October 26 I shared my takeaways from your comments about using AI tools while post-processing bird photographs. Your comments showed that photographers are much more likely to approve using these tools than non-photographers and I gave some general parameters to consider before sharing your photographs that have been modified using AI Generation on GreatBirdPics.com. On November 2nd I posed a rhetorical question: did being able to apply AI tools justify their use? On November 9 I had some fun and let Google’s Gemini read your comments and create its “interpretation” of what you said.
Today (just to prove I can beat a dead horse) I want to point out some final insights I’ve gained about using AI tools in post-processing and AI text-based tools.
Initially, I could have copy/pasted all of your comments into Gemini (without even reading them) and had it generate a synopsis of your thoughts. However I went “old school” and carefully read all your comments. As a result I gained a better understanding about your thoughts and formed my own opinions about where your thinking converged and diverged.
Had I initially just run your comments through Gemini and presented the summary it provided, would that have been considered deceptive to you the readers? Similarly is removing a branch in front of a bird or adding blue to a white sky deceptive to the viewer? At some level I believe it is.
I spent hours (25 revisions) writing the October 26 piece. One reason it took me so long was that as I thought through my responses, my opinions changed, which I believe yielded a better final version. Gemini took all of 2 seconds to digest your comments and to create a very nice summary. It used better vocabulary than I did, so the document doesn’t really sound like me. It would have been easy to use AI to write for me and I would have sounded much smarter than I really am, but I wouldn’t have learned as much .
It’s easy to use AI to improve my bird photographs. Does it make me lazy? Would I care less about composing my shots if I just thought, “I’ll take those branches out during post-processing”?
When I wrote the November 2nd post about justifying the use of AI, I gained more insight into my own thinking about bird photography and post-processing. Also, I don’t think AI could have come up with those same comments – I was searching my belief system about the ethics of using AI and I don’t think AI could have reproduced that thinking.
One more thought about using Generative AI. I went to a demonstration of how to use Lightroom in post-processing by Wildlife Photographer Joe Norton last week. During the demonstration he wanted to remove some branches and leaves from a photograph using the “old” Remove, Heal and Clone tools in Lightroom (with AI turned off). After a couple of minutes of fiddling around he was somewhat successful. After the presentation I asked him why he didn’t use the AI Generative Remove tool, which would have eliminated the offending objects better and in seconds. He responded, “They aren’t my pixels”. He went on to say that when you use Generative AI the highlighted areas of your photo are sent to the Lightroom Cloud and matched to other images and then sent back with its best guess as to how to replace the highlighted area. It doesn’t send back your pixels, but someone else’s pixels and places them in your photo. At first I scoffed at such hair-splitting: either mixing around the pixels in your photo to try to cover something or just grab the pixels that work best. I’ve pondered about “They aren’t my pixels” since then and I have a better understanding of what Joe was telling me; that when using Generative AI the photo isn’t completely yours anymore. I hadn’t thought about it that way before – thanks Joe!
If you enjoy seeing beautiful pictures of birds from around the world and reading about them Click Here to sign up for our mailing list. Members can post their own Great Bird Pics and learn more about bird photography techniques.