AI is quickly changing how we make Look-Up Tables (LUTs), and by 2026, it looks like anyone will be able to whip up a LUT from a reference image in no time. That sounds awesome, but there’s a lot more to it—especially if you’re working in Photoshop or ON1 Photo RAW. The problems aren’t just about AI. They’re baked into how LUTs work, what pros need for color grading, and where AI tech is right now.
Here’s what really matters: LUTs—Even AI Ones—Have Built-In Limits LUTs just aren’t content-aware. They’re like a giant spreadsheet of color values: you feed in some RGB numbers, it spits out new ones. That’s great for giving a photo a broad look—vintage, cinematic, matte, you name it. But they can’t look at an image and say, “Oh, let’s protect the highlights here, sharpen this part, or knock down the noise over there.” No matter how fancy the AI, an AI-generated LUT still can’t make those smart, local tweaks. You’re stuck with broad, global adjustments.
And that’s the point: LUTs hit the whole image at once. If you want to fine-tune specific parts of a picture, you’re out of luck.
Even if AI builds a LUT for you, you still need to know what you’re doing. Is your footage in log or Rec709? Use the wrong LUT on the wrong type, and suddenly your colors look awful, the dynamic range goes down the drain, and you get ugly banding. Some AI-generated LUTs might even try to include a log transform, and if you don’t apply it at 100%, you can break the color pipeline entirely. So, honestly, there’s always a learning curve.
Some workflows are even more unforgiving and you still need to understand the right way to use LUTs in your editing setup. Manual Grading Still Wins for Creative Control.
AI-generated LUTs give you a fast, custom starting point. But they’re just that—a starting point. But are LUTs ever more that that ? If you want complete creative control, nothing beats manual color grading.
When you do it by hand, you can adjust every color channel, curve, and parameter to nail the exact mood you’re after. AI LUTs? They’ll probably be more like fancy shortcuts or pre-made color presets, but they won’t replace the kind of detailed, personal touch you get from manual work. But is the totally manual method truly time and cost effective ?
So, the best workflow? Mix and match. Use AI-generated LUTs to get close, then jump in with manual adjustments to really make the image your own.
The REAL issue from User Experience : Naming and organizing LUTs. Built-in or Bought-in LUTs can be tough to figure out because they’re not always well-named or described. Imagine AI churning out dozens of custom LUTs—how do you know what each one does? Without clear names or descriptions, you could end up lost in a sea of mystery presets.
Compatibility Isn’t Optional. AI-generated LUTs need to come in standard formats, like .CUBE or .3DL, if you want to use them in ON1 or Photoshop. The tech seems on track to handle this, but it’s still a key requirement. No standard format, no dice.
Bottom line? AI is about to make LUT creation way faster and more flexible, especially if you want to match a certain look from a reference photo. But even the smartest AI can’t dodge the basic limits of LUTs. You still have to know how to use them, understand what they’re made for, and do some manual color grading if you want results that really stand out.