Free art? Maybe not

Recently, we see AI develops pushing out new platforms for using AI to create art. Examples include DALL-E and Stable Diffusion.

But are such platforms really free?

It is often said that if a service is free, then you are the product. In the case of such AI tools for art, is this also true?

Deep learning neural networks have brought about significant progress in image recognition technology. The move from recognising images to creating them is part of the natural progression. There is, however, a core problem for computers in the creation of images: what to create?

Computers are great (and getting even better) at doing what they are told to do. They can now automatically control a series of machines to produce a product. They can drive a car to reach a designated location. But they still lack that upstream process: What to make? Where to go?

In art, this is the same. What to draw?

So while the use of AI tools for creating art may be free to the user in terms of money, make no mistake: you are paying. With your ideas. The ideas of what to draw is being collected, which can then be used to further train AI in that aspect that they lack: what to draw.

And when AI eventually learns human creativity, what is going to set us apart from machines? What is the value of an artist's creativity and techniques, when an AI can do so much better, having been trained using the creativity of millions of artists and billions of art pieces?