Sarah Silverman sues OpenAI and Meta over copied memoir
Comedian and actor Sarah Silverman is suing OpenAI and Meta, alleging that the technology companies developed artificial intelligence tools that freely copied her memoir, "The Bedwetter," without permission.
Silverman, an Emmy-winning performer and former cast member on "Saturday Night Live," is the latest content creator to file a lawsuit over so-called large language models (LLM), which underpin burgeoning "generative" AI apps such as ChatGPT. LLMs develop their functionality by "training" on vast amounts of written and other content, including material created by professional and amateur writers.
Silverman's lawyers say training AI by having it process others' intellectual property, including copyrighted material like books, amounts to "grift." In parallel complaints filed July 7 along with two other authors, Chris Golden and Richard Kadrey, Silverman accused OpenAI — which created ChatGPT — and Facebook owner Meta of copying her work "without consent, without credit and without compensation." The plaintiffs are seeking injunctions to stop OpenAI and Meta from using the authors' works, as well as monetary damages.
In exhibits accompanying the complaints, filed in the U.S. District Court for the Northern District of California, ChatGPT is asked to summarize Silverman's memoir, as well as works by the other authors. It produces accurate summaries as well as passages lifted verbatim from the works, but doesn't include the copyright information that is customarily printed in these and other books — evidence that it was fed a complete copy of the work, according to the complaint.
OpenAI and Meta both trained their respective LLMs in part on "shadow libraries" — repositories of vast amounts of pirated books that are "flagrantly illegal," according to the plaintiffs' lawyers. Books provide a particularly valuable training material for generative AI tools because they "offer the best examples of high-quality longform writing," according to the complaint, citing internal research from OpenAI.
OpenAI and Meta did not immediately respond to a request for comment.
Joseph Saveri and Matthew Butterick, the attorneys representing the authors, in January also sued Stability AI on behalf of visual artists who accused the "parasite" app of glomming off their work. Last year the duo filed a lawsuit against GitHub, alleging its AI-assisted coding tool built on stolen coders' work.
The AI field is seeing a vast influx of money as investors position themselves for what's believed to be the next big thing in computing, but so far commercial applications of the technology has been hit or miss. Efforts to use generative AI to produce news articles have resulted in content riddled with basic errors and outright plagiarism. A lawyer using ChatGPT for court filings also was fined after the tool invented nonexistent cases to populate his briefs.