Richard van der Oost writes:
When we think of 3D rendering, we generally think of the creation of imagery that will eventually be consumed humans. (3D animation, VFX, etc.)
However, there is a whole industry that is starting to use 3D rendering to generate input data for AI and machine learning models. I.e. the rendered images will be consumed by computers/robots.
Read more in this article about how this is done, why it's done, and how you can get involved.
5 Comments
I want distance from iA... Who would be foolish to create Content to teach machines to steal work from artists in yet another area in the future...
They help... the increasingly rich AI creators with other people's work and the poorest artists
I read your article... it doesn't make any sense...
Why am I going to create models and invest my time in something that can make a living for artists in the future?
Thanks for the content, now I know how it works and how I can sabotage iAs by creating low quality content with wrong names to confuse and delay learning...
I couldn't do this alone, but imagine millions of artists from forums like Reddit? Imagine the damage ^_~
@YOUR BOSS,
I absolutely agree with you.
Unfortunately this is already happening on a large scale. There is a reason why a company like Epic Games now own Artstation (2D), Sketchfab (3D) and Bandcamp (music). They've already invested heavily in everything AI as a game company and now they'll be going for content. Despite many protests from the Artstation community, they have hardly changed their policy. So Artstation has been scraped for 2D content for training AI models without any consent from artists. AI-generated content is cheaper for game production so they couldn't care less about protests. Next in line is 3D, video and music. And when you read through a couple of research papers, you'll know this is already happening.
Artists should stop being naive about this and take action. In the least like you suggest, although it's probably too late already. Hopefully legislation will change things, but I'm not optimistic. The court cases against Stable Diffusion will be the litmus test and determine the future of many, many creators.
@TAURELIUS
Grateful for the support, let's stick together preserving what we love to do... True art...
Together we are stronger
Thanks for the feedback guys.
I agree with what you say.
However, when you read "AI", it doesn't automatically mean "AI for content/art creation".
This article covers how human generated 3D content can help AI do useful stuff.
For example: Look at camera footage of plants, and recognise when they need to be watered. (silly example) but it can be anything. Help self driving cars be more safe, you name it.
It's not about replacing human 3D art with AI generated art. But instead about replacing real-world photo's with 3D renders as it can be more efficient in this context. The same way car ads are often times 3D rendered instead of photographed. Hope that makes sense.
Maybe I was not fully clear about this distinction in the article.
Appreciate the engagement! cheers