Authors Lose AI Copyright Court Case in San Francisco

Well, this outcome was pretty much guaranteed. I don’t like these companies ripping off artists by using their art to train systems that use mass statistical analysis to come up with stuff that looks somewhat like it has been produced by artists. But:

1. The question about something being copyrightable is all about work. You have to show that what you produced entails original work. In our current understanding of this legal framework, you need to be a human to produce original work. Therefore, machine learning algorithms can’t produce anything that’s legally considered a work. Whatever they produce can’t be copyrighted, so how would the process they use to get there infrige anyone else’s copyright?
2. When these algorithms look at artists’ works to come up with something they themselves generate, they basically do the same thing a human artist does to come up with their art. A writer needs to read books to learn how to write and is inspired by other people’s work all the time. The same goes for visual art. Or music. How else do you explain this? It’s just that machines do this at scale — that’s just what machines do.
3. I personally think the idea that machine learning algorithms threaten artists is simply wrong. If that is the case either you’re not a good artist or your audience is dumb — in many cases maybe both. What we call AI has nothing to do with actual intelligence and it can’t be creative. It’s simply statistical analysis on a very large data set. Anything not in the dataset it can’t even imagine. Because it can’t technically imagine anything. We don’t even understand the process by which humans do this. All these systems can do is rehash old stuff and make it look good. That can be very useful in many cases. But it isn’t art and it does not compete with actual human creativity. Sure, it can replace a romance dime novel that is written purely to make the author money. But it will never be a Terry Pratchett, a Joseph Heller or a Franz Kafka.
4. And finally: If you declare that feeding text into a computer isn’t fair use under copyright law, you may just as well say goodbye to search engines or anything else that enables you to get anything useful out of the huge mountain of random data we call the internet.

Reuters: Meta fends off authors’ US copyright lawsuit over AI

A federal judge ruled on Wednesday for Meta Platforms against a group of authors who had argued that its use of their books without permission to train its artificial intelligence system infringed their copyrights. U.S. District Judge Vince Chhabria, in San Francisco, said in his decision that the authors had not presented enough evidence that Meta’s AI would dilute the market for their work to show that the company’s conduct was illegal under U.S. copyright law.

Chhabria also said, however, that using copyrighted work without permission to train AI would be unlawful in “many circumstances,” splitting with another federal judge in San Francisco who found on Monday¹ in a separate lawsuit that Anthropic’s AI training made “fair use” of copyrighted materials. “This ruling does not stand for the proposition that Meta’s use of copyrighted materials to train its language models is lawful,” Chhabria said. “It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one.”

AI companies argue their systems make fair use of copyrighted material by studying it to learn to create new, transformative content, and that being forced to pay copyright holders for their work could hamstring the burgeoning AI industry. Copyright owners say AI companies unlawfully copy their work to generate competing content that threatens their livelihoods.