There's much more in the article. If you're not a lawyer—especially if you're not a lawyer—I expect you'll get from it a better understanding of copyright and how it could be better. Thx for reading!

7/7

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6999539

Of the two threads, comparative criticism is far more pro-business. You can be a megacorp or a 10-year old kid and violate copyright under the compare-the-texts approach. That makes comparative criticism divisive, and if comparative criticism==copyright, then that makes copyright divisive.

But if our interest is in making sure that money is distributed fairly when there's money on the table, there's a century of copyright jurisprudence focusing on that question, not micromanaging artists.

6/7

The simpler approach to AI, I propose, is to go back to the other thread: when you say you're going to mass-harvest the work of creators so that you can replace those creators at mass scale, that's dead center in the sphere of unfair business dealings. If you only pay creators with enough leverage and tell the others you're entitled to ignore stated conditions about a copyrighted work and take whatever you want, that's unfair dealing.

Isn't that easy? You don't need metaphysics.

5/7

When somebody tells you that a work is transformative and therefore fair use, they're equating comparative criticism with the whole of copyright law.

Fediverse, the thing I most want you to get out of this paper is that the ordinary property approach to copyright is why copyright is problematic. We don't have to go full libertarian here and declare that anybody can take anybody's work, but there's a vision of copyright without everybody on the defensive about when imitation is too much.

4/7

The ordinary property approach leads to a sort of analysis where judges try to isolate the spark of creativity in an earlier work and determine whether a later work used that spark, and if so, whether it added sufficiently novel changes to that spark. It's comparative art criticism, but by judges.

Brought to AI, it turns into downright metaphysics: is the spark of human creativity encoded in model parameters? And now everybody is confused.

3/7

There are two threads of copyright law:
∙ As protection of unique artistic expressions. There is a "spark of creativity" which is like any ordinary property.
∙ Unauthorized copying as bad faith business dealings. Creators can't create knowing that anyone can copy their work for free.

Both are part of the judicial understanding of copyright law, but at different times in history, one has been more prevalent than the other. Right now, the first is very ascendant.

2/7

OK, intellectual property fans, here's a draft law review essay I wrote for you about AI and copyright, entitled "The art critics won't save you from the robots". It's really about a trend in copyright jurisprudence which is distorting the AI debate. You shift the trend, you fix the debate—and possibly a lot of what's wrong with copyright.

1/7

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6999539

(5a/5) Me: I'm distributing my latest book, _Topology: The Interesting Parts_ , via Bandcamp.
It's a math book, intended as a more recreational book for people who don't read math books.

https://b--k.bandcamp.com/album/topology-the-interesting-parts

Mastodon, is there some kind of dentistry kink community I don't know about? Because I'm having a hard time convincing the insurance company that my getting a root canal retreated is out of medical necessity and not some kind of recreational activity.

My #introduction on another site: I'm a microeconomist who spends his days searching for better models, but I mostly use this site for one-liners.

Am a heavy music consumer, and may post track links w/background. I do a lot with intellectual property, migration, and computational statstics, which I think are important topics with a lot of room for curiosity and creative exploration. I make soap, bike everywhere, often sleep in tents, and am generally an introverted hippie.