The Art of Taking

Artists have always borrowed.

That statement is sometimes offered as a defense of generative AI.

And there is truth in it.

A painter studies other painters.

A musician learns songs written by other musicians.

A novelist absorbs thousands of sentences before writing one of her own.

Apprentices once learned by copying masters.

Artistic movements exist precisely because ideas move between people.

Impressionism influenced post-impressionism.

Blues influenced jazz and rock.

African artistic traditions profoundly influenced European modernism.

Sampling became a creative language of its own in hip-hop.

Culture grows through transmission.

There may be no such thing as a completely uninfluenced artist, but influence has never meant that everything is permissible.

We recognize the differences between learning and copying, between homage and imitation, between inspiration and appropriation.

Those boundaries have always been imperfect.

Artificial intelligence did not create that ambiguity.

It changed its scale.

A human artist might spend years studying another artist's work, absorbing certain techniques while developing a style of her own.

A generative system can process relationships across enormous collections of human-created material and allow someone who has never written song lyrics to type a few words and produce something resembling a recognizable artistic style in seconds.

The person making the request may never know who created the works that helped make that capability possible.

The original artist may never know either.

That distance changes something.

Before AI, copying often required enough effort that we were aware we were copying.

Now we can simply prompt.