Thoughts on AI and Composite Photography

Do you use AI or composites?

No.  My gallery images are digital photography captures from real moments in time.  While I do spend an extensive amount of time editing/processing all images before adding the final product to my respective galleries, I will never replace a sky, change clouds, or otherwise replace, add, or subtract any major elements that were not present in the scene at the time.  For example, I will never add a moon or sun rays to an image if they weren’t there when I was.  I believe that doing so more closely resembles digital art and is not staying true to the craft of landscape photography. 

I also believe that other photographers who add or subtract major elements of a scene using composites or AI should offer full disclosure so the general public can understand what they are looking at.  Lately, on social media, I have come across Lake Tahoe landscape images that were created using AI and were shared/posted with zero disclosure. Many viewers initially failed to recognize how these images were created and that it was not an actual photograph of a real place. But locals, who are familiar with the Lake Tahoe shoreline, quickly figured it out and explained in the comment section that it was not a real location and that the image had to be AI generated. I would think that this was an embarrassing moment for the “photographer.”

In summary, I do not believe that composites or AI have a legitimate place in landscape photography. However, it’s use in product photography, as a marketing tool, or in other types of photography to fit a specific need for a client is understandable and inevitable. Regardless, I think disclosure is key if these techniques are implemented and the results aren’t obvious to most viewers. 

What about High Dynamic Range?

The above controversial techniques (composite photography, AI digital art) should not be confused with HDR where multiple images are shot at different exposures and blended together to enhance the tonal range of a final image.  While the vast majority of the images in my galleries are single exposures, there are times where I will combine multiple exposures into one. I believe that this is a completely legitimate photographic technique that can enhance an image if it is done properly. At the same time, I often see highly unrealistic looking images in landscape photography where I think the this technique was pushed too far, but that is ultimately just a matter of taste and preference.

AI Editing Tools

One non controversial aspect of AI that is helpful for photographers are the editing tools that can speed up and streamline the post processing workload. Suddenly, thanks to AI, it recently became easier to create masks by selecting a sky or subject with the click of a button in Adobe Lightroom and Photoshop. While these masking tools are sometimes imperfect and need to be fined tuned for landscape photography, they are already very helpful for portrait photography and I am confident that they will just get better over time. 

Future of AI in Landscape Photography

While AI photography technology will inevitably just get better and better, nothing will replace the experience of capturing a real moment in nature. Fortunately, I am confident that there will always be a market for authenticity in landscape/adventure photography and I don’t expect AI to replace that.

UPDATES 1.28.24

I won’t share a screenshot of the image, but the most recent viral AI landscape photograph I came across on Facebook is a perfect example of how ridiculous and problematic this trend can be. When I first saw this Southern Utah image with a rainbow and three different colored lightning bolts at different elevations that were so clearly AI generated or just recklessly copied and pasted into the scene in the most unrealistic manner, I found it mildly amusing. But, then I noticed it has over 30K “likes” and a sea of comments complimenting the “photographer” for an amazing capture. There was no disclosure that the image was fake and a surprisingly large number of people couldn’t tell the difference. Strange times …