this is a bit reductionist
some people spend hours tweaking the same prompt to get exactly what they're picturing into the image
some will use inpainting to carefully and selectively replace parts of the image until it fixes things that look imperfect
...
but then you get into semantics about the amount of time taken to create "art"...is photography an art? isn't it possible to get lucky and snap something in 5 seconds that becomes iconic throughout history?
Everyone will have to excuse the wall of text but the response turned out to be longer than expected
Whether someone is using AI or creating it themselves the result it still artwork. However a defining characteristic of an artist is what they're bringing to it, how substantive their input in the result is. Otherwise they're essentially directing the AI like a client.
If we consider sampling in music where controllers are used, they don't typically refer to themselves as musicians even if they're using instrumentation since they recognize they're utilizing others' work. Exceptions would be for example where they're creating original drum sequences or substantially altering the samples. They even have to get approval for the samples and pay royalties to the original artists (even if some skirt this).
Masking objects is in itself is a skill, which AI masking for inpainting is a simpler form of. Simpler since it's both not the same isolation purpose as general masking for compositing and since AI front-ends offer cruder tools than one would be accomplishing natively in image editors, outside of Photoshop-integrated Stable Diffusion plugins.
Similarly regular inpainting is a skill, which again the AI form is a vastly simplified and automated version of (in regular inpainting one is more manually extending the image behind masked out objects, while AI fully automates this and integrates arbitrary additions). Yet being good at either or both on their own don't arguably make one an artist, even in professional settings.
Some people are just getting a taste of the difficulties and skill of such things which others do more rigorously and all the time (but don't generally post about) and thinking those aspects mean they should be called the artist. And that's only for those in your selective example who are going to more effort, most (from the many examples and breakdowns I've seen) are only engaging with crude AI masking/inpainting and spend the bulk of the time tailoring the prompts, which themselves by necessity compromise of predominantly keywords/tokens of other artists' names and tags the dataset were trained with (models also have different limits to the number of tokens they accept, the rest get ignored).
If they want to call themselves prompt engineers at least it's closer to an accurate term. They can feel the reward to spend what they feel is a long time finding the right set of tokens and seeds to get an essentially magic black box to pull an image out of noise, maybe mask the hands to roll the dice on non-garbled anatomy, mask the background for more inpainting prompts, but most are ultimately just guiding its direction rather than creating the result.
If they get to the point where they're isolating individual elements and compositing separately in their own creative manner then they're at least engaging in a more artistic process. If they're transforming an illustration of their own making using img2img (more than just guided by an MS Paint-esque doodle), to where the output is directly influenced by it then they're likely already an artist utilizing AI for iteration rather than relying on it.
Concept artists and collage artists are among those who already utilize even direct sources but substantially transform them as part of a larger whole that they're applying their artistic skills to (illustration, conceptualization, compositing/masking/layering, etc). To me that's the crucial aspect: whoever is behind the substantial element of the result. If it's the AI, then the AI is the artist.
