Last year I spent $20 on Fiverr to get a background removed from a headshot. The turnaround was two days. The result was fine. Then someone mentioned remove.bg, and I did the same job in about four seconds for free.
That was the moment I realized the errand was gone. Not simplified, not cheaper. Gone.
I've been comfortable with image tools for a long time. I know my way around Photoshop. I use Canva regularly. I'm not a designer, but I've never been the type to panic when someone hands me a PSD file. So when AI started showing up inside the tools I already used, I didn't have a dramatic conversion moment. It was more like the water got warmer without anyone turning the dial.
Background removal is basically solved
This is the clearest win. Tools like remove.bg, PhotoRoom, and Clipdrop have turned what used to be a careful, tedious masking job into something you don't even think about anymore. Hair edges, transparent objects, low-contrast scenes. I still remember spending real time with the pen tool on those kinds of cutouts. Now I upload the image, and it's done before I've finished reaching for my coffee.
PhotoRoom has become especially useful for anyone doing ecommerce or product content. It handles the removal and then lets you drop the subject onto AI-generated backgrounds with matched lighting and shadows. Remove.bg is still my go-to when I just need a clean cutout and nothing else, because it does that one thing extremely well and doesn't try to upsell me on a full editing suite while I'm there.
For most content work, any of these will get you there. The differences only show up on the hard stuff like wispy hair or semi-transparent fabric. For a blog featured image or a social post, it genuinely does not matter which one you pick.
Typing your way through an edit
The stranger shift is generative editing. Nano Banana, Google's image model inside Gemini, lets you upload a photo and describe what you want changed in plain English. "Remove the person in the background." "Change the wall color to blue." "Add rain to the window." You type it like you're leaving a note for a retoucher, and the AI just does it.
I've used Photoshop's Generative Fill for similar tasks, and it's impressive in a different way. It's surgical. You select an area, describe what should go there, and it fills it in with something that usually looks convincing at normal viewing distance. For fixing a distracting element in a photo or extending a background to fit a wider crop, it saves real time.
But here's the thing I keep coming back to. When I use these tools, the work doesn't feel like editing anymore. It feels like directing. I'm not making decisions with my hands the way I do when I'm adjusting curves or painting a mask. I'm writing short instructions and evaluating the output. Which is, honestly, closer to what I do all day as a writer anyway. Maybe that's why it feels so natural, and also why it feels slightly off. The skill I spent years building in Photoshop, the muscle memory around selections and layers and blend modes, just matters less now for a lot of the tasks I used to use it for.
The upscaling and cleanup side
Then there's the restoration and enhancement category. Topaz Photo AI is the standout here. Feed it a noisy, soft, or low-resolution image and it will denoise, sharpen, and upscale it in ways that genuinely look good. I've rescued client photos that I would have told them to reshoot two years ago. Luminar Neo does similar work with a broader creative toolkit on top, things like sky replacement and automatic portrait retouching that lean more toward stylistic choices than corrections.
These tools sit in a different part of the workflow than the generative stuff. They're fixing problems rather than inventing content. And for that reason, I trust them more instinctively. When Topaz sharpens a blurry photo, I can see what it did and judge whether it worked. When Nano Banana adds rain to a window that never had rain, I'm relying on the AI's taste as much as my own.
The part nobody talks about
Here's what actually changed for me, and I don't see this discussed enough. Because I can handle all of this myself now, I'm expected to. One client I work with used to have an in-house person who'd take my article text and build the featured image. At some point last year they stopped doing that and started asking me to supply finished images with the draft. No conversation about it. No adjustment to the rate. The capability just silently became part of the job description.
I'm not bitter about it. The tools make it fast enough that it doesn't wreck my day. But I do think writers and content people should pay attention to this pattern of expanding expectations. Every tool that makes something easy also makes it expected. That's not a reason to avoid the tools. It is a reason to think about where your rate reflects the full scope of what you're delivering.
Where I've actually landed
I use remove.bg or PhotoRoom weekly. Canva's AI features have replaced a lot of what I used to open Photoshop for. I reach for Nano Banana when I want to experiment with something visual and I'm not sure what I want yet. And I still open Photoshop when the job actually requires precision, because nothing else gives you that level of control when you need it.
The overall feeling is positive, with a caveat. These tools made me faster and more self-sufficient, and I like both of those things. But they also compressed a set of skills that used to take years to develop into something anyone can do with a text prompt. If you're someone who built value around being the person in the room who could handle images, that compression is worth sitting with for a minute.
Has your job quietly absorbed skills you never originally signed up for?
Top comments (0)