NO VISUALS TODAY, JUST FOOD FOR THOUGHT.
Dear people, there is this thing we need to discuss. A dilemma, for some it is one of a moral nature about the artistic means of expression, and for others it’s all good. We have been thinking about sharing with you a few very interesting prompts we have stumbled upon recently related to street photography. At first we were like, OMG, this is so good. How great is MJ at imitating the very unique style of WordPress award wining photos, or the widely-recognised style of Magnum photographers. It’s nearly as good as the original Bang Bang Club work. If you can prompt a street shot like this, then the sky is the limit, like literally. Prompt the shot, upscale, work a bit on its final outcome in PS and there you go, Alex Webb himself, almost. There is no way anybody can tell the difference. Therefore, it also means anybody can mimic the unique style of… anyone. Is that what we are in need for? Are we ok with it? Aren’t we in a kind of loop here? We have started asking ourselves those question the very moment Nick St. Pierre (@nickfloats) started asking his:
‘I don’t consider generating an image using AI to be stealing per se. But if you’re using an artist’s name in your prompt you are most definitely stealing IMO. It’s one thing to use an artist’s name for experimentation, and if you openly share your prompts it’s not quite as bad. But if you’re regularly using artist names to achieve a certain style (their style) and posting that work without sharing the prompt, that’s a big no-no for me.’
– Nick St. Pierre*
On one hand, we would say hell yes! He is right, get your own style, how could you be riding on the renowned artists’ coattails? On the other, well, aren’t we all, as humankind like that? Isn’t it natural?
The AI is theoretically doing what every human does when they evaluate the works of others before crafting their own style in any artistic endeavor. In fact, you could argue a human is „stealing” more given their output is based on a small sample size of inspirational artists and worldly experience compared to LLMs trained on the entire internet taking inspiration from billions of people.
If so, then what is the difference in using an artist name, or describing the desired style while prompting? And how about mixing styles of many artist in one single prompt? Aren’t we, people the actual tools in the whole creation process, and the AI the real creator? Wasn’t the whole AI concept build on the right of free expression, for all?! Is it hurting anybody? There’s so many questions here, but really no black and white answers we could give, nor can anybody IMO.
The only two things by far we could do is either watch the advancement from the side of the road or ride it like you stole it 😉 just teasing here! But really, it is the audience that calls the shots here. A similar discussion has been taking place in the music industry for years but sampling related. Time will tell if it goes the same way. “Imitation is the sincerest form of flattery” goes the old saying, and finally one of the greatest, Pablo Picasso supposedly said “Good artists copy; great artists steal.”