At a recent cross country meet, two of my Nikon cameras shot the same runners in the same morning light. My D5 gave me skin tones I loved. My Z9 gave me pinkish faces and orange arms. So I gave ChatGPT and Claude my photos and my camera settings to see if AI could help me fix it, both in Lightroom and in the camera for next time. It didn't fix everything. It did find a setting I didn't know was off.
Watch it here, then keep reading for the short version and a free download.
We delivered over 6,400 images from that meet. I edit every frame, but I want to do as little as possible, so I shoot JPEG at big events. With JPEG, what comes out of the camera is most of the final look. The more I get right in camera, the less I fix later.
I gave both AIs three straight-out-of-camera JPEGs from each camera, plus screenshots of the full settings for each one. I pulled those from NX Studio, Nikon's free software. Lightroom shows your exposure, but not color settings like white balance or Picture Control, and those mattered most here.
One tip: use the full file info, not the summary. My summary panel showed a different white balance setting than the full info did. Give AI the wrong settings and you get the wrong answer.
Both AIs pointed to Picture Control, the camera's built-in recipe for color, contrast and sharpness in your JPEGs. I thought my two cameras were set the same. They weren't. Both said Standard, but my Z9 had saturation at +1.00 and contrast at +1.50. My D5 was at +0.50 on both.
A setting with the same name isn't the same recipe on every camera. Both AIs also pointed out that different cameras and lenses won't give you identical color, even with matching settings.
ChatGPT gave me a step-by-step plan with starting numbers: lower contrast and highlights, open the shadows a little, and make small cuts to clarity and saturation. Then it surprised me. Leave white balance alone and go after the skin with Point Color, a Lightroom tool where you pick one exact color, like skin, and adjust only that. That step made the biggest difference.
Claude adjusted more of the overall image, including temperature and tint, and leaned on the color mixer and masking.
The two results were very similar. Both took out most of the pink. Neither got all the way to my D5. The blacks weren't as crisp and the clarity wasn't there. But both beat the edits I'd tried on my own.
You don't need two cameras for this. Any photo you like can be your goal image, even one of your own from a good day. The free versions of both tools are a fine place to start.
The most common mistake is changing everything the AI suggests at once. Change one thing at a time, so you know what actually fixed it.
Free guide: AI Prompts for Photographers
The exact prompts I used, with fill-in blanks for your camera, plus a one-camera version. Sign up at the link below and I'll email it to you, along with new guides as I make them.
Get the free guideWould you fix skin tones in camera, in editing, or both? Have you asked AI for help with settings or editing? Leave it in the comments on the video. I read all of them.
Get out there and make beautiful images.