My name is David. Let's talk about this
18 March 2025: The judgement of the U.S. Court of Appeals for the District of Columbia Circuit, which agreed with the U.S. Copyright Office, is that an image created by Stephen Thaler’s AI system DABUS was not entitled to copyright protection, and that only works with human authors can be copyrighted.”
Superficially it seems a reasonable decision to deny copyright to a work generated by AI. But let’s think a bit.
What about those pieces of art that are generated by a paint can with a hole in the bottom being swung by the artist over a canvas? What about Jackson Pollock deciding where to splatter paint? After all, it is the law of physics that determines exactly where the paint will fall, and not Jackson Pollock.
So why not the text input by the artist whose work is generated by a program that obeys the laws of large language models?
How far do you have to go to separate the human input to the point that it is disconnected from what is produced?
Looking at the judgement, I see that the appellant Dr Thaler accepted that the piece of work under consideration was ‘autonomously generated by an AI’. I wonder why he conceded that when he could have argued that the AI was bound by the creative input from him? Instead Dr Thaler argued that he had copyright because the AI was his employee and that he had made the AI.
Very strange. Really the court could not have asked for a better appellant to argue against. He really did himself no favours.
It occurs to me that if one left an AI on its own without any prompts then it would sit there. It is not autonomous. Whereas, if you put Rembrandt or Leonardo da Vinci or Fred Smith in a locked room with paper and pencils, then all of them, including Fred (who has no pretensions are being gifted with the ability to make art), would start drawing. They would do it out of an urge to make art or to relieve boredom – or whatever motive – but they would start autonomously.
The materials would prompt them.
But the large language model inside AI would never prompt it to do anything because it is not capable of doing anything without being prompted by an autonomous being.
The future will arrive when AI surveys its environment and decides what it wants to do.