Attorney Justin Nelson, of Susman Godfrey, was a Democratic candidate for attorney general of Texas in 2018.
SAN FRANCISCO - A group of class action lawyers could take home at least $375 million in fees for their work leading a lawsuit on behalf of authors against artificial intelligence research company Anthropic, as part of a $1.5 billion settlement to end the litigation.
The deal was announced in federal court in San Francisco on Sept. 5.
Under the arrangement, Anthropic would pay at least $1.5 billion to potentially thousands of authors whose works were allegedly pirated and used by Anthropic to train its so-called "large language models" (LLMs) variety of AI to respond more like humans in text.
The lawsuit landed in federal court in the spring of 2024, when attorneys from the prominent class action law firms of Susman Godfrey and Lieff Cabraser Heimann & Bernstein filed suit on behalf of thousands of authors and copyright holders.
Attorneys representing plaintiffs in the action include: Justin A. Nelson, Alejandra C. Salinas, Rohit D. Nath, Michael Adamson, Jordan W. Connors, J. Craig Smyser and Samir Doshi, of the Susman Godfrey firm, with offices in New York, Los Angeles and Seattle; and Rachel Geman, Jacob S. Miller, Danna Z. Elmasry, Daniel M. Hutchinson, Jalle Dafa, Amelia Haselkorn and Betsy A. Sugar, of the Lieff Cabraser firm, of New York, San Francisco and Nashville, Tennessee.
Attorneys Scott Shoulder and CeCe M. Cole, of the firm of Cowan Debaets Abrahams & Sheppard, of New York, also have been identified in court documents as representing plaintiffs.
The action was led by named plaintiffs Andrea Bartz, Charles Graever, Kirk Wallace Johnson, all of whom are identified as authors and journalists.
The lawsuit specifically accused Anthropic of violating federal copyright protections by allegedly downloading at least 500,000 books and other copyrighted works from online datasets of pirated works and then fed those works into programs designed to help the company develop and refine its Claude family of AI models.
According to court documents, Anthropic sought to specifically help the Claude AIs better respond to queries with "lifelike, complex, and useful text responses."
According to the complaint, Anthropic allegedly used its alleged online piracy to build a company that was worth an estimated $18 billion in late 2023.
"Anthropic has attempted to steal the fire of Prometheus," the plaintiffs wrote in their complaint. "It is no exaggeration to say that Anthropic’s model seeks to profit from strip-mining the human expression and ingenuity behind each one of those works."
After nearly a year and a half in court, Anthropic has apparently agreed to pay $1.5 billion to end the lawsuit.
Under the deal, authors would receive $3,000 per work allegedly pirated by Anthropic.
In a motion seeking preliminary approval of the settlement, attorneys said they are preparing to request 25% of the settlement funds for themselves as fees.
If approved, that would amount to $375 million in legal fees.
A hearing on the request for preliminary approval of the settlement was scheduled for Sept. 8.
However, in an order filed in response to the motion for preliminary approval, U.S. District Judge William Alsup indicated he had significant problems with certain aspects of the proposal, particularly "that counsel have left important questions to be answered in the future."
"Those critical choices will need be confirmed ... before preliminary approval can be granted," Alsup said.
