IP Law Daily, COPYRIGHT—S.D.N.Y.: Former class plaintiffs avoid being pulled back into AI cases they brought, (Oct 9, 2025)

Law Firms Mentioned:Latham & Watkins LLP

Organizations Mentioned:Latham & Watkins, LLP | Lieff Cabraser Heimann & Bernstein, LLP | Microsoft | OpenAI Inc.

By Matthew Hersh, J.D.

The five writers were once named plaintiffs in different cases against OpenAI, but dropped out once the cases were consolidated.

Sarah Silverman, Ta-Nehisi Coates, and three other writers will not be deposed in ongoing class-action litigation against ...

By Matthew Hersh, J.D.

The five writers were once named plaintiffs in different cases against OpenAI, but dropped out once the cases were consolidated.

Sarah Silverman, Ta-Nehisi Coates, and three other writers will not be deposed in ongoing class-action litigation against OpenAI, the federal court for Manhattan has held. The court, in the latest of a flurry of discovery rulings in a series of now-consolidated actions against the generative artificial intelligence company, found that the mere fact that the five writers were once named plaintiffs in other lawsuits against OpenAI, without more, could support their being dragged back into the case (In Re: OpenAI, Inc., Copyright Infringement Litigation, No. 1:23-cv-08292-SHS-OTW (S.D.N.Y. Oct. 8, 2025)).

The decision is the latest step in far-ranging class action litigation against OpenAI for alleged infringement of copyrighted works via its highly popular chatbot. Most of that litigation has been focused in either the Northern District of California or the Southern District of New York. Named plaintiffs in those cases have included writers such as John Grisham, Sarah Silverman, David Baldacci, Jodi Picoult, Paul Tremblay, Richard Kadrey, Michael Chabon, Ta-Nehisi Coates, along with organizations such as the Authors Guild.

The unique posture of this particular dispute arises out of the April 2025 consolidation of those cases—four in California, eight in New York—into a single case before longtime Judge Sidney H. Stein of the Southern District. When that happened, several of the named plaintiffs in the original cases were no longer named in the consolidated class action complaint. OpenAI moved to depose some now-absent plaintiffs Douglas Preston, Simon Winchester, Ta-Nehisi Coates, Sarah Silverman, and Junot Díaz, leading to this opinion.

Depositions of absent class plaintiffs. The court denied the motion. OpenAI contended that the five witnesses would have relevant testimony both as to its fair use defense as well as to the question of certification. But discovery from absent class members, the court noted, “runs the risk of turning the ‘opt-out’ class action into an impermissible opt-in class action.” Accordingly, the court determined, OpenAI could go forward with this discovery only with a “strong showing” that “(1) the discovery is not sought for any improper purpose, to harass, or to alter the membership of the class; (2) it is plainly relevant for purposes at trial or issues pertaining to class certification; and (3) the discovery is narrowly tailored, does not impose an undue burden on absent class members, and is not available from a party.” Under this standard, the court found, none of the depositions made the grade.

Simon Winchester: OpenAI contended that he was the only person among the current and former named plaintiffs who opted into a licensing agreement between HarperCollins and Microsoft concerning generative AI. Thus, OpenAI argued, only he could speak to how the agreement worked from the “perspective of a putative class member.” But “OpenAI does not explain whether they have sought such discovery from Microsoft and, if so, what information about the licensing agreement they have not received from Microsoft that can only be obtained from Mr. Winchester,” the court found.
Douglas Preston. OpenAI asserted that, as a former president of the Authors Guild, he had a “beneficial financial relationship” with OpenAI. His testimony, OpenAI argued, could therefore be used to argue that there a conflict among the putative members of the class. But just because a proposed class “received any benefit that can be linked in any way to a defendant,” the court noted, does not mean that “such a tangentially-related benefit is strong evidence of a conflict among the class and is thus relevant.” Here, OpenAI did not contend that Preston was “receiving a benefit directly from OpenAI’s alleged copyright violations,” the court noted, and whether he did or did not have a beneficial financial relationship with OpenAI was “too attenuated to justify compelling the deposition of an absent class member.”
Ta-Nehisi Coates: OpenAI sought to depose him about certain statements that he made about the impact of generative artificial intelligence that were allegedly relevant to OpenAI’s fair use defense and class certification. But these statements, the court noted, did not in any way appear to relate to OpenAI or the large language models that power it—they were instead generalized statements about the impact of AI-generated art. This case, by contrast, the court noted, “concerns copyrighted books and articles and text outputs of Defendants’ LLMs, not AI-generated art.” Statements about a defendant’s “general industry,” the court found, were simply “too attenuated” to support a deposition here.
Sarah Silverman. OpenAI sought to depose her regarding public statements she has allegedly made about this and other lawsuits against GenAI companies. OpenAI contended that she has given multiple interviews explaining “her views on OpenAI, ChatGPT, and [GenAI]” and that her statements allegedly “undercut” the class plaintiffs’ arguments about whether causation and harm are susceptible to common proof. But as with the statements by Coates, the court noted, “generalized statements about GenAI or the GenAI industry are not relevant to fair use.” Moreover, the court noted, OpenAI failed to identify “what information Ms. Silverman would testify about that is unique to her and unavailable from any of the named class plaintiffs.”
Junot Díaz. Finally, OpenAI sought to depose him regarding his testimony in a separate lawsuit about his knowledge of monetary harm arising from OpenAI’s alleged copyright infringement. OpenAI alleges these statements contradict the class plaintiffs’ claims of market harm and are relevant to class certification. But “the testimony in question goes, at best, to Mr. Díaz’s state of mind regarding what harm, if any, he may have suffered at the time he was deposed” in that lawsuit, the court found. “OpenAI fails to explain why Mr. Díaz’s state of mind testimony is so unique and unavailable from other named class plaintiffs such that a deposition of an absent class member is appropriate and narrowly tailored.”

The Case is No. 1:23-cv-08292-SHS-OTW.

Judge: Wang, O.

Attorneys: Anna Josefine Freymann (Lieff Cabraser Heimann & Bernstein, LLP) for Authors Guild. Allison S. Blanco (Latham & Watkins LLP) for OpenAI Inc.

Companies: OpenAI Inc.

Cases: Copyright TechnologyInternet NewYorkNews AINews

IP Law Daily, COPYRIGHT—S.D.N.Y.: Former class plaintiffs avoid being pulled back into AI cases they brought, (Oct 9, 2025)

Law Firms Mentioned:Latham & Watkins LLP

Organizations Mentioned:Latham & Watkins, LLP | Lieff Cabraser Heimann & Bernstein, LLP | Microsoft | OpenAI Inc.

By Matthew Hersh, J.D.

The five writers were once named plaintiffs in different cases against OpenAI, but dropped out once the cases were consolidated.

Sarah Silverman, Ta-Nehisi Coates, and three other writers will not be deposed in ongoing class-action litigation against ...

By Matthew Hersh, J.D.

The five writers were once named plaintiffs in different cases against OpenAI, but dropped out once the cases were consolidated.

Simon Winchester: OpenAI contended that he was the only person among the current and former named plaintiffs who opted into a licensing agreement between HarperCollins and Microsoft concerning generative AI. Thus, OpenAI argued, only he could speak to how the agreement worked from the “perspective of a putative class member.” But “OpenAI does not explain whether they have sought such discovery from Microsoft and, if so, what information about the licensing agreement they have not received from Microsoft that can only be obtained from Mr. Winchester,” the court found.
Douglas Preston. OpenAI asserted that, as a former president of the Authors Guild, he had a “beneficial financial relationship” with OpenAI. His testimony, OpenAI argued, could therefore be used to argue that there a conflict among the putative members of the class. But just because a proposed class “received any benefit that can be linked in any way to a defendant,” the court noted, does not mean that “such a tangentially-related benefit is strong evidence of a conflict among the class and is thus relevant.” Here, OpenAI did not contend that Preston was “receiving a benefit directly from OpenAI’s alleged copyright violations,” the court noted, and whether he did or did not have a beneficial financial relationship with OpenAI was “too attenuated to justify compelling the deposition of an absent class member.”
Ta-Nehisi Coates: OpenAI sought to depose him about certain statements that he made about the impact of generative artificial intelligence that were allegedly relevant to OpenAI’s fair use defense and class certification. But these statements, the court noted, did not in any way appear to relate to OpenAI or the large language models that power it—they were instead generalized statements about the impact of AI-generated art. This case, by contrast, the court noted, “concerns copyrighted books and articles and text outputs of Defendants’ LLMs, not AI-generated art.” Statements about a defendant’s “general industry,” the court found, were simply “too attenuated” to support a deposition here.
Sarah Silverman. OpenAI sought to depose her regarding public statements she has allegedly made about this and other lawsuits against GenAI companies. OpenAI contended that she has given multiple interviews explaining “her views on OpenAI, ChatGPT, and [GenAI]” and that her statements allegedly “undercut” the class plaintiffs’ arguments about whether causation and harm are susceptible to common proof. But as with the statements by Coates, the court noted, “generalized statements about GenAI or the GenAI industry are not relevant to fair use.” Moreover, the court noted, OpenAI failed to identify “what information Ms. Silverman would testify about that is unique to her and unavailable from any of the named class plaintiffs.”
Junot Díaz. Finally, OpenAI sought to depose him regarding his testimony in a separate lawsuit about his knowledge of monetary harm arising from OpenAI’s alleged copyright infringement. OpenAI alleges these statements contradict the class plaintiffs’ claims of market harm and are relevant to class certification. But “the testimony in question goes, at best, to Mr. Díaz’s state of mind regarding what harm, if any, he may have suffered at the time he was deposed” in that lawsuit, the court found. “OpenAI fails to explain why Mr. Díaz’s state of mind testimony is so unique and unavailable from other named class plaintiffs such that a deposition of an absent class member is appropriate and narrowly tailored.”

The Case is No. 1:23-cv-08292-SHS-OTW.

Judge: Wang, O.

Attorneys: Anna Josefine Freymann (Lieff Cabraser Heimann & Bernstein, LLP) for Authors Guild. Allison S. Blanco (Latham & Watkins LLP) for OpenAI Inc.

Companies: OpenAI Inc.