Jo, if all you were doing was playing with ideas, creating, experimenting, it's a boon to you. But it's a threat to the original artists. 200 years from now that won't matter. While those artists and their families are trying to live on what they produce today, it's a serious problem.
The example used here is Erin Hanson, whose artworks keep grabbing me on Pinterest. I can tell her style (and her usual subjects) instantly. She not only works hard to produce constantly, she also has (or partners with?) a few live galleries to make her living. They literally stole from her directly in my view without her permission and the AI works are too close to hers to be anything but confusing to the market in copyright terms.
Yes, the government can do more about this via copyright laws, but the burden is on the artist/owner of the copyright to take it to court.
You know parody most of the time when you see it. You know derivative works when you see them pretty much always. But this is sooooo close to home that it's not just uncanny, it's unconscionable to distribute (much more so to sell) them.
When David Hockney started doing his work on an iPad, that was just experimentation or legitimate original use of a new medium. AI copies are not, even if initiated by a flesh and blood human.
There is, alas, no way to copyright a
style of painting, and there is nothing much anyone can do about any new technology other than trying to insulate himself. As it stands, Erin Hanson is, at least for the moment, thoroughly AI-proof, for the simple reason that she sells originals, painted in physical media. That is not something AI can do at all. The technology might wreak havoc in the print market, but then again, maybe not as much as we fear.
For one thing, the technology isn't mature, as amazing as the images sometimes look at first glance. I saw this one "inspired by Hopper" posted on Twitter:
Quite impressive, but look carefully. Even impressionistically painted figures should make some sense, and many of these do not. Look particularly at the one on the far right. What exactly is going on there? Not to mention the text on the image: "CAFE CAFE"? "NEW YOOK"? And I see this all the time in AI art; there is almost always something weird or "off". It remains to be seen whether they'll be able to fix this.
Now if you want prints for your home or office, and AI ones are much cheaper than ones made after the work of humans, then perhaps it will be worth it for some. But I think many people might prefer human art on principle; having your home or office full of AI prints might come to be seen as a sign of lack of sophistication.
Well, maybe, maybe not. No one can predict the future, or what a generation of people who grow up with AI will prefer. At least for the moment, while I'm pretty impressed with the technology, and I have seen some genuinely beautiful and striking imagery made with it, I keep on asking myself: what is it actually good for? Thus far it is not of much use to illustrate books, for example, it is of no use for people who want original paintings on their walls, it has limited use as posters and such things as album covers.
I think at least some low end illustration jobs are going to go away, and one will have to see what effect this has on the market. I don't think it is likely that human art will go away, any more than AI made human chess players go away. But it is possible that it will make human
professional art, i.e. the possibility of making a living through art, go away. Well, for the vast bulk of human history there was no such thing as a professional artist, so perhaps we should be philosophical about it. But as I note above, I don't think the danger is immediate or catastrophic.
Also, at least some forms of human art is very AI-proof indeed. One I already mentioned: physical paintings. Another that occurred to me: comic books and graphic novels. Go page through, say, one of the later Tintin books, and ask yourself what the chances are that any conceivable AI will be able to come up with such a thing.
Lastly, just as a point of interest, something I saw someone note on Twitter. We all predicted AI completely wrongly. In the arts, we always thought that should AI ever be able to produce art, it will tend to be coldly accurate but soulless art. The exact opposite happened: AI images are often strikingly beautiful, but upon closer examination full of amusing errors, accuracy-wise.
Similarly, we thought AI would make "low end" human jobs go away. The exact opposite happened: we are going to have AI lawyers and psychologists long before we have AI janitors. This fits in with something else I saw someone mention. AI can nowadays also write texts, to some extent. One guy tried this: for a speech, he made some bullet points and asked an AI writer to expand these into a long text. The AI had no problem doing this, and came up with a quite impressive text that needed virtually no correction.
But then the guy did the opposite: he gave the AI a long text and asked it to summarize it into bullet points. The AI failed to do this, completely and utterly. This follows because AIs don't actually understand anything. But it does show how we humans are very easily impressed by color and noise. The stuff politicians have been saying for years in their speeches is pretty much the same kind of stuff the AI above came up with when it expanded the bullet points. Utter drivel, and we mindlessly cheer it on.
It may be that we need to rethink our aesthetics, rather than rant at AI for being able to so easily emulate it.
And yes, after a long absence from CS, I am back for a while.
