AI and the future of art

nutanc
Ozonetel AI
Published in
2 min readSep 18, 2022

--

As we all know, there has been an explosion of AI generated art, thanks to models like stable diffusion(which is awesome BTW). But the ownership of this art is still a grey area. Lot’s of debate on Twitter and other forums. Mostly the two sides of the coin seem to be:

1. AI generated art is fine. What if it is trained on other people’s art work. Even people get inspired from other art works.

2. AI generated art is wrong because permission was not taken from the creators.

Both seem valid arguments. But an argument seeming valid is not enough. Is there any legal precedence? What is legal?

Also there are too many grey areas. What if an artist has created some digital art and put it up for sale. Can I use that art as the prompt(img2img) and generate similar art and also put up that new art for sale?

For example, here, I have created AI art similar to famous paintings, https://medium.com/ozonetel-ai/stable-diffusion-painting-most-famous-paintings-64b2117dc8ff

Personally, I think the “AI generated art is fine” gang misses the forest for the trees. It is not about “how” the AI has learnt something and is reproducing stuff or whether it is copying. It’s about whether the artist has “licensed” their work to be used for training. Just because the technology has evolved so fast that the legalese for this has not been invented yet, does not mean the artists lose out. So if an AI is allowed to do this, are humans also allowed to do this?

So what is going to be the future of art. Will AI take away the jobs of artists?

If I had to make a prediction, I predict the following:

New licenses will be defined. Artists will chose those licenses and decide if they want to allow AI to be trained on their art or not. Just like robots.txt allows websites to be crawled or not.

Because the licenses and rules don’t exist currently, businesses have been training AI to create art. But here is the problem. The AI has copied only the current art styles. It can generate only current art styles.

Soon, these styles will become cheap to produce and not in demand anymore.

Artists will create new styles and block AI from learning their styles. Which will allow their styles to be more in demand.

So, my prediction is that soon, we will have two types of art. Machine generated cheap art and human generated costly art.

--

--

nutanc
Ozonetel AI

Love software and love building software products. Blog at http://t.co/yICZVFPd