“James Cuda, the CEO of Procreate […] expressing his dislike of generative AI and explaining that Procreate’s products are designed with the intention that ‘a human will be creating something.’ The announcement appears to have resonated with many digital artists who have expressed concerns about the integration of AI into creative tools.”
7 Likes
A good reason to buy Procreate if I needed another graphic app!
Mark
4 Likes
In the short term, it’s a great marketing gambit which plays to their target audience. In the long term, it raises some interesting questions about drawing and computers.
When an artist uses Procreate on an ipad, it creates an illusion like they are drawing the same way they do with a pencil and paper. Yet it is only an illusion. The Apple Pencil doesn’t make a mark. Rubbing the glass on an iPad doesn’t make a mark, either. There’s actually no mark being made at all, in the way that a pencil marks a sketchpad.
Instead, the artist is using a complex device (the Apple Pencil) to transmit multi-channel data driven by the angle, pressure and grip on the pencil to the computer in the iPad. The software is following complex instructions to interpret these signals and load them into a memory buffer, and then a graphics subsystem is reading that buffer and turning screen pixels on and off. This happens so fast that it seems like the pencil is causing the mark to appear on the screen, but it isn’t. The computer is turning lights on and off to make you think there is a mark.
Procreate also has artist assist features like snapping to perspective lines and automatic line straightening. This takes the imprecise gesture from the user, and creates a better version in the computer, and loads that in memory and to the screen.
It raises the question: at what point does the automation cease to help the artist, or become objectionable non-art? Sometime soon, AI will be able to watch what an unskilled artist tries to draw, and will offer an improved replacement. At what point does this become anti-artist? I’m legitimately asking the question, because it seems to me that if you’re artisitically comfortable with telling an iPad what you want drawn, then having AI help you draw above your ability isn’t such a big stretch.
What do you think about this?
1 Like
Good question. I don’t pretend to know the definite answer. The medium (device) in itself isn’t a problem in my book. Otherwise e.g. every writer had to use pen (or typewriter) and paper to be a “real writer”. Which makes no sense.
Those “assist” features are already murky water. But then, is a spell checker considered cheating? It helps people to write above their orthographic abilities. But so do human editors. It gets even murkier when it comes to developmental editors…
Telling someone (or something) to draw or write or compose an artwork for you that you couldn’t even hope to create in your wildest dreams? Yeah. Probably safe to say: That’s beyond the red line.
You could argue that a movie is also a collaborative effort, no matter how talented the writer or director is. It usually takes a lot of people and technology and money to make it happen. But the audience expects that. And there’s a long disclaimer end credits honoring everyone involved.
Maybe all “AI” art needs a mandatory disclaimer and end credits. Right now it doesn’t, because you can tell. But when that stuff starts to remember the correct amount of fingers…
It gets even fuzzier when you look at arts like photography, which has never been about the artist physically making lines. Darkroom manipulation has always intervened between what was captured on film and what is shown to the audience. Digital “enhancements” to focus, exposure, contrast, color, etc. are now available to anyone with a smartphone, and are being driven by more and more sophisticated algorithms. What level of manipulation is “allowed” before the final image ceases to be a “human creation?”
1 Like
I’m not sure I’d agree. I can program a drum pattern in Garage Band that I can’t play on the drums. I don’t think that makes it any less my creation.
Of course, all of this is subjective. There are people who – to this day – believe that Bob Dylan is no longer a “real musician” because he started using an electric guitar in 1965.
As far as that goes, I guess my line in the sand is how much influence the creator has over the creation. If writing a novel is like ordering a pizza, IMHO it’s not really a human creation. If the process is more like being a sous-chef, making corrections and changes as need be, I think I’m more comfortable calling that a human creation.
Of course, this whole area is the battleground of the future of creative media. We live in interesting times.
2 Likes
If you program it, it’s your creation (unless you claim you played it on the drums). But if you tell an “AI” to program it for you…
I admit that there’s a huge gray area in the middle and I can’t tell exactly how far it extends.
But I know for certain that I don’t want to look (listen, read) at nightmare-inducing “artworks” anymore. It’s something about those apparently dead things trying to look real. Probably Uncannyvalleyphobia or something.
Yeah, that’s the rub, isn’t it? And the grey area is likely to shift with global attitudes towards AI.
There was a time when using a section of another musician’s recording in your own was considered a creative crime. Now sampling is completely ordinary, and the question is whether you use free samples or pay for famous ones – but using them is completely normalized.
It won’t surprise me when AI goes the same way.
3 Likes
The same argument can be applied to anybody who uses a computer/device for anything related to “creativity”.
So the same can be applied to writers of novels & stories, poetry, plays, film scripts etc. I do hope you’re not using Scrivener for creativity, because, by your own comments, you aren’t actually creating anything.
That’s quite the leap of logic and not at all a natural extension of what @popcornflix was saying.
5 Likes
I agree with you - you’re just restating my ipad/procreate argument. I’m perfectly comfortable with the idea of the ipad and procreate helping an artist with their draftsmanship. I think I’m more comfortable than most with the idea of AI-assisted art being legitimate art.
It also brings to mind the professional musicians who create “generative music” using no AI. They use complex electronic switches and purpose-built computers to run programs and algorithms to create unique machine-generated music. The human composer who builds the programs and wires up the circuits owns the copyright to the music, and this has been going on for decades.
It makes you wonder where the line between “art” and “non-art” should be drawn.
1 Like
Well, I guess I read your intentions wrong then. In that case… I think you’re both (presumably deliberately?) taking an extreme exaggerated position which is (in my view) not supportable / defendable.
So…
Actually, you can’t. You can move some sliders that change the emphasis, but it’s actually pre-created drum patterns and beats in GarageBand. The song you use it in might be your creation (thanks to the terms of the software legal agreement), but (thanks to some interesting court cases, including one where the drummer from The Smiths successfully sued Morrissey and Marr) the drum part most definitely isn’t your creation in this example. It would be if you were actually programming beats in a drum machine, but GarageBand’s Drummer tool doesn’t work like that.
Yes, but if you’re going to deputise people into your argument who are clearly delusional you’re going to come across as delusional yourself.
I agree.
This is interesting. I kind of but not completely disagree with your example… If you’re the Executive Chef, coming up with the recipe, and telling the sous-chef what to cook, and the sous-chef occasionally makes a suggestion that might improve the dish, some of which you accept and some you don’t, then I’m comfortable calling the recipe the Executive Chef’s creation but the meal the Sous-Chef’s creation.
In music terms, this is the difference between publishing rights in a song, and ownership of the recordings (ignoring the third category of performing rights for this example). This difference is why Taylor Swift can re-record her old albums, and also why the music industry is a little more comfortable with assistive creation tools.
Most art, including novel writing, painting, photography etc) doesn’t have an equivalent split baked in, which is part of the reason why an obvious answer on where the line is drawn is hard to intuit.
Great example. But remember, although it’s normalised, you still have to pay the original producer of the sample and get their permission. If you don’t, you will lose in court big styleee. Ask Richard Ashcroft how much money he made from The Verve’s Bittersweet Symphony, for example! (spoiler alert, LEGALLY, he got nothing, with 100% of the royalties going to the Rolling Stones. Jagger and Richards decades later gifted the song back to Ashcroft – very nice of them).
(as another fun aside, Vanilla Ice is an absolute legend for his approach to being sued for Ice Ice Baby – when Queen sued him for ripping off Under Pressure (noting that this wasn’t a sample, but an inspired-by near copy of the riff), Vanilla Ice cheaply bought the publishing rights to Under Pressure from under their noses and made an absolute fortune off both songs!).
Here’s where you lose me.
In my mind (and this is just my opinion, not a legal or cultural fact), the question of creation comes down to decision making. The procreate line fixing is just the digital equivalent of using a ruler – it’s not making a creative decision. If you ask Adobe Photoshop to generate an image based on a prompt, you’ve not passed the threshold of creation (in my view), and calling yourself an artist just because you picked which of three pictures it generated you wanted to use would be sad and pathetic. Lumping all digital tools, including writing in Scrivener, as equivalent is just being deliberately provocative.
(which I admire, obviously 
)
2 Likes
Let me clarify -
I think copyrightable creation using AI is a continuum. On one extreme, giving minimal input to the AI (“make a picture of a boat on a lake”) should warrant no copyright protection, and on the other extreme, painting a picture mostly by hand but using AI-generated images like a collage artist would use stock photos should be protectable and owned by the human artist.
Somewhere along that continuum, we each subjectively decide where we think the AI is doing too much for it to be considered human-created art and thereby protectable. Each of us will pick a different point on the range. The key is finding tolerable compromise for black letter law. The compromise will be inherently arbitrary, but creating a bright-line standard is essential in order to include AI into commercial art.
Before AI, the USCO ruled that a photograph could be copyrighted because the photographer made creative choices regarding composition, lighting, and timing, even though the camera mechanically captured the image.
I believe the same standard should apply to AI - if your input to the system specified content, composition, lighting and rendering, it should be protectable by the artist. In the same way, if a writer created an outline and then had an AI write the prose to turn the outline into a book, I believe that author should own the copyright.
Your opinion may differ, and that’s okay by me.
P.S. - Garage Band has a step sequencer for the drums called Beat Sequencer, so you can start from scratch and program a drum pattern.
2 Likes
May I just say that, as I draft an inevitably unsatisfactory reply to the NaNoWriMo/AI debacle thread, what a pleasure it is to read a nuanced discussion of these issues on our forums. Unfortunately my mind is so frazzled that I have nothing to add. I have no idea where the line should be drawn either, but the points raised in the posts here are fascinating.
As you were.
8 Likes