Ultimately, I think the best solution is thinking in terms of readiness for truly sophont AI. Current AI barely resembles the AI of sci-fi in terms of the actual processing going on, even if "AI" can fuction at a sufficient level to make someone unaware think they are just a bit odd.
But, sooner or later we will develop truly sophont AI, and that will create socio-economic issues that will be of a scale and nature comparable to the Atlantic slave trade.
In my view, it is obvious that today's "AI" is just a marketing-speak buzzword for clever algorithms that can broadly mimic intelligence. But, by creating a legal framework now where these AI products are considered entities in their own right independent of their creator organisations, we can shape a much more cooperative social relationship to true AI when it emerges.
As far as the usage of these tools goes?
Pragmatically, we would all like to have pretty pictures in our AARs, and I believe we can all at least have a vision of what we have in mind, even if we may not necessarily be able to put it on a page. In an ideal world those of us without sufficient artistic talent would pay artists with it, but paying artists is tricky, as any artist will tell you. For those of us on a budget of shoestrings, if the choice is between AI art, and no pictures...
I totally understand both sides.
In an ideal world, AI corporations would properly obtain consent for usage of content, and that they generally don't is what creates the primary ethical issue with using these tools; the process by which the AI develops it's pictures and the process by which we put together separate concepts into one concept are analogous in outcome even if the processes are dissimilar.
For me?
I could commission an artist to draw exactly what I have in mind if I describe it adequately at a substantial outlay per picture - the artist has every right to charge what they charge of course, no dispute there.
But if a single portrait is $45 say, that compares to almost 4 months of a subscription to Playground AI, which is the AI art generator I use, and pay for because I tested it out and concluded it was good enough for practical purposes (
some prompts you'll get something usable every time you click generate, others... Well, it took literally hundreds of pictures to get acceptable images for Naomi and Buri for Life2.0, and the software is not capable of generating images of nuclear pulse propulsion utilising vessels) and as
@Chac1 has pointed out, the delay you get as a free user is extremely limiting. (
which I do understand; they have to prioritise paying users after all)
On balance though, a Playground subscription is far better value for money.