In an article I wrote for Publishers Weekly late last year, I wrote that Generative Artificial Intelligence ("GAI") challenges accepted notions of creativity and authorship.
I also wrote that to understand copyright law, you had to understand the policy behind copyright law. The premise is that without copyright, authors would have no incentive to create new works.
However, algorithms and artificial intelligence don't require incentives in the same way humans do. Unlike book publishers, who generate royalties for human authors, internet platforms prioritize data-driven and machine-learning engagement for advertising revenue. In doing so, they harness user interactions and behavior to sustain their financial models. As a result, these AI systems can generate vast amounts of content, from good enough to outright toxic, blending fact and fiction without any regard for copyright protections or permissions. This glut of AI-generated media poses complex questions about information quality and attribution as well as the boundaries of creativity and originality.
As the volume of AI-generated media increases, the provenance of information will become more important, creating market incentives and consumer demand for publishers and creators who can demonstrate authenticity and high quality.While addressing the complexities of regulating AI-generated content remains an open question, the established community of publishers has an important advantage in addressing the comfort level of consumers as provenance plays a central role in fostering trust and reliability in information. Publishers (with a capital “P”), through selectivity in what they acquire, careful editing, collaboration amongst sales and marketing, publicity, and the payment of royalties, offer a baseline of trust in the data they publish.
Without trademarks, John Oathout, author of Trademarks, wrote, "consumers would have no basis for selection or rejection, or any assurance that a particular product is the product they are seeking."
 |
| Alfred A. Knopf |
Alfred A. Knopf (the man, not the imprint) was keenly aware of that proposition when he wrote The Borzoi Credo, a publishing manifesto that appeared in the November 1957 issue of The Atlantic Monthly.
It read, in part, "I believe that a publisher's imprint means something and that if readers paid more attention to the publisher of the books they buy, their chances of being disappointed would be infinitely less."
Unlike copyright law, trademark law can be used to stop the unauthorized use of a bestselling author's name, a series title, symbols, and markings that the public associates with a particular publisher or other source. In this respect, trademark law is an effective cudgel against those who pass off their wares as endorsed by or coming from an established creator, publisher, or producer.While the publishing industry understandably has antagonism towards large language models, the industry will no doubt take an active part in shaping the future of AI, whether through legislation, licensing their books to train AI, creating bespoke AI models with their own curated datasets, and trumpeting the Good Housekeeping Seal of Approval value of their author and publisher brands.
To return to the premise of this post, that GAI is undermining the incentives given to authors by copyright law while fulfilling the purpose of copyright, that tension will work itself out over time, but we need human editors and publishers for transparency, accountability, and quality control purposes.
