Why the generative AI stuff?

[Reevyv]

Why do you use generative AI for the preset pack images?

One of them even has the watermark from one of the stolen artworks the ai was trained on on them (Or the “K” is for “Earthquake Bass”, the name of the pack)

It also looks so cheap. I used to love Arturia for their UI and overall design as well as attention to detail in this stuff but I don’t really want to support a company that sells tools for creatives (musicians) but profits from the exploitation of other creatives (visual artists). And it’s so unnecessary too. If you can’t afford a graphic designer for a 15€ preset pack, just put a plain/empty image there. Would be better than this.

Luckily there are a lot of developers that don’t do this although it’s sad to see Arturia go this route. Maybe I’ll get back to Arturia products in the future if they stop this nonsense.

Sorry for this rant but I know some actual artists so I’m biased in this topic and see it’s relevance…
If Arturia sees this: The question at the top was an honest question. Was it just for saving money? Was it laziness?

Also, the intention of this topic is not to be hate speech or anything. This is just my honest opinion and view on this. Sorry if this doesn’t come across as intended.

[Lowkus]

AI builds a statistical understanding of objects. If you were to show it pictures of frogs, it will see the typical shapes and features of frogs, so when it’s asked to draw a frog it’ll be like “Oh yeah, that’s one of the green chubby things with bulbous eyes and it’s got the haunchy little legs and no tail”. It’s basically learning as a human would, but it has better memory and precision. Unfortunately, AI is still kind of stupid about certain things, like when it sees watermarks in pictures it can start to think that maybe it should add watermarks to what it creates, and then it’ll scribble in an abstract looking thing that’s kind of like a watermark.

AI is not unique in using other artists’ work to learn. Most artists learn that way. In fact, many professional artists create their art by looking at photographs and translating them into drawings. The vast majority never get the photographer’s permission, so it seems hypocritical for artists to complain about how AI is doing exactly the same thing. My suspicion is that artists don’t truly care about AI having learned from their art, rather they are grabbing at any excuse to shut down AI out of fear of the competition and loss of income. But while artists might lose income due to AI, everyone else on the planet is likely benefiting from AI by having an almost free robot artist at their disposal. Though honestly I don’t think AI will ever fully replace human artists, but it will definitely have a negative impact on artist incomes overall. Frankly, artists have had a cushy lifestyle in that they’ve been able to earn money while doing what they love. Rather than fighting against AI art, they might just need to do what the rest of us have done in life… get a job that isn’t necessarily what we love but manages to pay the bills.

[Reevyv]

Hey @Lowkus .

The difference is that real art has a sense. Art emerges when an artist (human) deals with themselves, their environment or something entirely different. Like a dialogue, but the language being music, a painting or a poem for example.
You can apply the same thing to the process of getting inspiration or learning an artistic skill. Artists learn with sense, by actively paying attention to some specific detail in a painting. Or a producer that listens to a song and really likes how the synth lead runs through a phaser effect in the brigde part (random example).
The personality of the artist is an essential factor. Not just in the process of making art, but also while learning/getting inspired.

AI has just masses of stolen art in a database and guesses what fits logically to the given prompt. There is no sense, understanding and awareness at all. Neither in the processes, nor in the product.

You can tell by the watermark that it has no sense and that the art in the database wasn’t used to learn but was just sort of mushed together while trying to resemble something.

Artists also don’t have the ambition to replace one another but rather to develop a unique personal style (or sound, for musicians)

AI developers also act really disrespectful towards artists. I bet most artists are willing to tell you their greatest inspirations, because it matters to them. While the AI developers don’t even know what they’re consuming.
When you go into the cinema to watch a movie and listen to the score, that’s fine. But it’s not okay to record the movie and to harm the film makers by stealing their content. That is essentially what all the AI companies are doing.

I hope you understood some of my arguments, have a great day.

[Lowkus]

AI art has more human involvement than a lot of people give it credit for. You can have human choices in designing a workflow to make the AI engine (if it’s being run locally in something like ComfyUI for example, my own workflow took about four weeks of work to create). Then the human writes a prompt which is really a pure act of creativity, and while it might sound easy to write a prompt there are people who rewrite the prompt over and over and over as they try to massage it in convoluted ways until the AI is forced to generate images closer to what the human wants. Then, after the AI has churned out images, the human has to comb through them all to choose the “best”, which is another artistic choice. In some cases the human will take an additional step where they retouch the image to further make it look exactly as they want. I wouldn’t say AI image generation is more difficult than other arts, but I believe it represents enough human involvement to consider it a form of art. I’d probably rank it similar to street photography, where the photographer does not force their creative impulses into the work as much as they observe what comes along and they make the decision of when to press a button that captures what they like.