The Scholarly Kitchen has a useful discussion of the state of play on the vexed question of whether a work produced by a computer can be protected by copyright. Thus far the thin red line is being held, and only works created by human beings can be copyrighted: animals, plants and machines are out of luck. After all if every possible combination of words were to be copyrightable, this would seem to mean the end of copyright, since anyone (merely human) writing anything would have to acknowledge that it could theoretically already have been written by some busy AI program.*

Of course we don’t yet have our Library of Babel, and until a “work” is recorded in some tangible form, it can’t be copyright — so if you are composing a magnificent verse epic in your head as you do your ten-mile runs, write the damn thing down before some machine beats you to it. Verse seems to be the area in which AI is currently performing best: lots of associative adjectives, and an ability to disguise a lack of logical argument. (See also Robot writer.) However I did write last year about Springer’s publishing of a machine generated survey of recent research. It’s not clear from the accounts about this volume whether Springer, or an “author”, is claiming copyright in this work. Probably not too important as this is not the sort of stuff you’d want to pirate.

I bet what’ll happen will be that machines will busily write away and some person or other will step forward to claim that they in fact wrote the book, for which they’ll then take out copyright. After all, even today your computer provides you will spellcheck and word prompts. Nobody thinks that these things represents a diminution of authorial creativity, so why not the next step, and then the next step, and so on. As long as you claim you wrote it yourself not your iPhone, who’s to know any better?

Of course come to think of it, in a Library of Babel world where everything is already available what is the value of something called copyright whose main job is to encourage writing?


* This puts me in mind of the recent story about a couple of guys who have written down every melody there could ever be.

Many musicians have been sued to the tune of millions of dollars, over melodies they wrote which courts have decided originated in a previous song. Damien Riehl and Noah Rubin have programmed up a way to make every melody ever written, or not yet written, available in the public domain. It’s all just mathematics after all: there’s a finite number of possible combinations of eight notes, and maybe creating a melody is nothing more than spiritually accessing a pre-exisiting database of all possible melodies.

If you don’t see a video here, please click on the title of this post in order to view it in your browser. Thanks to David Crotty at The Scholarly Kitchen for the link.

But will this work? Chris Meadows at TeleRead doubts it. Of course, as Mr Riehl admits copyright suits can be about more than just melodies, but some have been about melody alone, and to the extent that his hard drive does represent a copyrightable work it may reduce legal traffic. But does the collection of melodies not fail because of the ruling that only works created by humans may be copyrighted? Takes a law suit to discover what’s legal.