In October, I participated in a delightful conference at Columbia’s Kernochan Center on “Copyright Outside the Box.” I was on a panel dedicated to computer-generated works, along with my friends Bruce Boyden and Annemarie Bridy. Our respective symposium essays have just been published in the Columbia Journal of Law and the Arts, along with a slew of other fun papers.
While this post contains some obligatory self-promotion, it also features some discretionary cross-promotion. Read my essay, yes, but read it together with Bruce’s and Annemarie’s. While the three of us came at the topic and prepared our remarks independently, when we arrived that conference, it turned out that we had three complementary takes. The resulting papers read unusually well together.
The best starting point is Annemarie’s “The Evolution of Authorship: Work Made by Code,” which builds on her earlier “Coding Creativity: Copyright and the Artificially Intelligent Author.” She gives a delightful and evocative survey of the kinds of creativity on display in works created by or with the assistance of computer programs. This isn’t a 21st-century problem, she shows; it has been haunting computer labs and the Copyright Office for almost as long as there have been digital computers. She then makes the pragmatic point that we don’t need to resolve the philosophical questions of whether computers could count as people or be creative (whatever that means) in order to fit these works inside the copyright system. Instead, we can use existing devices like giving ownership of works made for hire to the hiring party to get sensible results.
Next, read my piece, “There’s No Such Thing as a Computer-Authored Work–And It’s a Good Thing, Too.” I make the stronger argument that regarding computers as authors is affirmatively wrong: it’s a bad solution to a non-problem. I take the argument that there’s nothing interestingly new here even further back: to creative machines like Spirographs and Musikalisches Würfelspiele. Computers aren’t in this respect different in kind from previous systems that sequenced creativity between the initial designer of a process and someone who carries it out. Along the way, I do some demystification about computer-assisted creativity; working digitally makes almost no difference to the question of what counts as authorship. There is no simple answer to how we should assign copyright ownership of computer-generated works: sometimes it’s the user and sometimes it’s the programmer and sometimes it’s both. But treating “the computer” as the author doesn’t make those hard questions any easier; in fact, it tends to obscure the important distinctions that matter in particular cases.
Finally, read Bruce’s “Emergent Works,” the paper I wish I’d written for the symposium. Bruce gives the best rejoinder to my argument I can think of: that what is potentially new with computers is their ability to yield “works that consist largely of creative elements that have emerged unbidden from the operation of the program.” These works are interesting because it may not be possible to “predict the work’s content with reasonable specificity before it is rendered or received by the user.” When neither the programmer nor the user of a program can make such predictions, he argues, the possibility of “emergent authorship” is worth taking seriously. This don’t have to mean treating computers as authors in the same we treat people as authors–and Bruce agrees with my claim that there is no one right answer. But he goes further in arguing that current doctrine may not always have a good answer, either. I’ve tried not to make my own piece inconsistent with Bruce’s argument, but he pinpoints the spot beyond which my argument gives out–which turns out to be along the curve of increasingly complex work-emitting programs Annemarie describes.
Download them all now: three for the price of one.