The gap between what took an hour in Photoshop and what takes a few seconds now is genuinely large — object removal, sky replacement, and noise reduction that used to require real manual skill are increasingly one click away. What I'd push back on is the framing that this makes editing skill obsolete. The tools are faster; the judgment about what to fix, and how much, hasn't gotten any easier to automate.
What AI Actually Changed
The honest summary is that AI collapsed the mechanical parts of editing — the masking, the cloning, the frame-by-frame noise reduction — into something close to instant, while leaving the judgment parts exactly where they were. Adobe's Generative Fill can extend a canvas or remove an object in seconds where a manual clone-stamp repair used to take real patience. DxO's DeepPRIME does noise reduction on high-ISO RAW files well enough to rescue shots that would previously have been unusable. Both are real, meaningful improvements. Neither one tells you whether an object should be removed, whether a sky replacement serves the image, or how much skin smoothing crosses from correction into distortion — that's still entirely on you.
Picking Tools Without Overspending
I wouldn't chase every AI-branded tool that gets marketed as revolutionary. For most people, the practical toolkit is smaller than it looks: something for RAW processing and general adjustments (Lightroom is the default most people should start with, and its free mobile tier covers more than beginners expect), something for noise reduction if you shoot in low light often enough that it matters (this is where DxO's DeepPRIME and similar dedicated tools earn their keep — general editors don't do noise reduction nearly as well), and a web-based background or object remover for quick, disposable edits where a full desktop workflow is overkill. Beyond that, tools like Luminar's sky and portrait AI features are genuinely useful for landscape and portrait work specifically, but I'd treat them as optional rather than essential — you can get most of the same result with more deliberate manual work in whatever editor you already own.
On subscriptions specifically: it's easy to end up paying for four or five overlapping tools that each do one AI trick well. Before adding a new one, I'd check whether the editor you already use has quietly added the same feature — Adobe, Lightroom, and most competitors have been shipping AI-assisted masking and selection tools into their existing apps rather than keeping them exclusive, so the tool you already pay for may already do what you're about to buy something new for.
Where Judgment Still Matters Most
Generative fill and object removal work best when what you're removing is simple and the surrounding area gives the algorithm enough context to convincingly fill the gap — a stray person at the edge of a landscape, a power line against open sky. They get noticeably less reliable the more complex the scene or the more specific what you're adding needs to be; a vague prompt asking the tool to "add something interesting" produces exactly the generic, slightly-off result you'd expect. I'd treat generative additions to a scene with real skepticism for anything beyond casual use, since they're the editing choice most likely to tip into fabricating content that wasn't actually there.
Portrait retouching is the other place where restraint matters more than the tool. Skin smoothing, eye brightening, and similar fixes can look natural in small doses and uncanny fast — the tell is usually skin that's lost all texture, which reads as artificial even to viewers who can't articulate why. Body reshaping tools exist and work technically, but I'd treat them as an ethical line worth being deliberate about rather than a default enhancement, especially for anything published without the subject's knowledge of what was changed.
A Rough Line on Ethics
The distinction I'd draw isn't "AI vs. manual" — a manual dodge-and-burn pass and an AI-powered one raise exactly the same ethical questions if the underlying choice is the same. The real line is between correction and fabrication. Color correction, removing a genuine distraction, reducing noise, and reasonable skin retouching are corrections — they clean up how a real scene was captured. Adding elements that weren't there, materially changing what an image documents, or editing photojournalism or documentary work in ways that misrepresent what happened cross into fabrication, and that's true whether it took an AI five seconds or a skilled retoucher an hour. If you're delivering client work, disclosing what was substantively changed — not every minor exposure tweak, but anything that alters what the image actually shows — is worth doing as a matter of trust, not just professional courtesy.
When I'd Reach for AI vs. Do It by Hand
For quick social posts and casual work, I'd lean on AI tools without much hesitation — the stakes are low and the time saved is real. For client deliverables, I'd use AI for the first pass and then review it manually, since generative fixes occasionally introduce subtle artifacts that are easy to miss at a glance but obvious at full resolution. For fine art or personal creative work, I'd use AI tools as a starting point at most — the whole point of that kind of photography is the specific creative choices you make, and outsourcing them to an algorithm's best guess defeats the purpose. Batch processing across a large shoot is the one place where AI is close to unambiguously the right call — reviewing hundreds of frames individually for adjustments that a smart preset can apply consistently is a poor use of time.
The Honest Bottom Line
The best edit is still one that serves the image rather than announcing itself. AI tools make it faster to get there, and in a few areas — noise reduction on difficult RAW files especially — they do things that were genuinely hard to do well by hand at all. But the eye for what an image needs, and the restraint to stop before an edit looks processed, hasn't been automated, and I don't think it's going to be. That part is still entirely the photographer's job.