By Terri Lively
Last year, Jane Friedman, author of "The Business of Being a Writer" realized a reader bought one of her books…but she didn’t write it. Friedman suspects that AI did.
Friedman's situation highlights a new problem writers face today—the threat to their livelihood and reputations posed by generative AI technology.
Generative artificial intelligence (AI) leverages algorithms to create new content, from audio clips to computer code to email drafts, among others. McKinsey & Company, a multinational strategy and management consulting company, predicts that generative AI applications could contribute up to $4.4 trillion to the global economy in coming years.
Generative AI software trains using previously written works. At first, the human supervised. However, developers eventually fed the AI models mass amounts of text—like 45 terabytes of text data—and the models trained themselves.
Is it Yours?
So, is some of this previously written work training generative AI yours? Maybe, maybe not.
Most authors don't want AI training with their work without their authorization. In a recent survey, an overwhelming 96 percent of authors said that a writer's consent should be required and that they expect compensation.
So, What Can You Do About It?
You can contact the publisher if you suspect that work is published under your name that you didn't write. For Amazon, Kindle Direct Publishing provides a link to an infringement report. Goodreads has you join the Goodreads Librarians Group where you can file a "book issue."
It's less clear what you can do about AI training by using your published work. The Author's Guild, a nonprofit organization representing working writers, wants compensation for you.
In July 2023, the Author's Guild met with lawmakers to encourage legislation that protects writers from being exploited by and, worse, replaced by AI. The Author's Guild is also working with Amazon to get fake books removed. The writer's advocacy group also calls for AI companies to let authors say whether AI models can use their work for training and transparency that identifies AI-generated text for readers.
There is litigation in the works, too, which could help. The Joseph Saveri Law Firm has a Generative AI Litigation page that invites authors to contact them with their concerns. The law firm filed a case last June, Infringement of Book Author's Copyrights, in the US District Court for the Northern District of California, seeking compensation from OpenAI, the creators of ChatGPT, for copywriting infringement and other grievances, including violations of California's unfair competition laws. Some high-profile authors have joined the case as plaintiffs, like Sarah Silverman and Paul Tremblay, claiming that the software used pirated copies of their works in training.
New technology often creates new problems for specific industries. This time, writers get to face them. If McKinsey and Company is right, plenty of revenue should be available to compensate today's working writers. Whether writers will see any is up to the courts and lawmakers…or, of course, the good conscience of the technology companies.
So, maybe don't hold your breath waiting for a check.
Comments