Judge Bars Authors From Relying on AI 'Dilution' Study in OpenAI Copyright Case

Did a federal judge decide whether AI-generated books dilute the market for human authors? No. On October 6, 2026, U.S. District Judge Sidney H. Stein granted, without opposition, a motion by OpenAI and Microsoft in In re OpenAI, Inc. Copyright Infringement Litigation, No. 25-md-3143 (SHS) (OTW), in the Southern District of New York.

The order says: "The motion by defendants OpenAI and Microsoft to strike the Lasinski Supplemental Report and to preclude Class Plaintiffs from citing or relying upon the paper by Class Plaintiffs' expert Dr. Tuhin Chakrabarty entitled, "Generative AI Floods and Dilutes the Market for Books" (the "Chakrabarty Paper") is granted without opposition." It cites docket numbers 2045 and 2087. The order relates to Nos. 1:23-cv-08292 and 1:23-cv-10211.

The order does not say whether generative AI floods or dilutes the book market. It does not rule on fair use. It strikes one supplemental report and bars the class plaintiffs from citing or relying on this paper.

The motion

On September 23, 2026, OpenAI and Microsoft noticed a "Motion to Strike Susman Godfrey L.L.P.-Funded Study and Supplemental Report of Michael Lasinski," under Federal Rules of Civil Procedure 26 and 37. Keker, Van Nest & Peters represents OpenAI. Orrick, Herrington & Sutcliffe represents Microsoft. A footnote says the motion does not address admissibility under Federal Rule of Evidence 702 or Daubert v. Merrell Dow Pharmaceuticals. The companies said they would raise any such challenge separately.

What the companies argued

This is the companies' argument in the September 23 memorandum, not a finding by the court. They argue that class counsel Susman Godfrey paid Stony Brook University professor Dr. Tuhin Chakrabarty to research how AI-generated books dilute the market for human authors, and that he co-authored a working, non-peer-reviewed paper. The memorandum says that on July 22, 2026, after expert-report deadlines, Dr. Chakrabarty, Dr. Paramveer Dhillon, Xinyue Liu, and Columbia Law School professor Jane Ginsburg uploaded a paper titled "Generative AI floods and dilutes the market for books." Later versions followed on July 26 and August 3. The companies say it remains a "Working Paper Under Review."

They say a version of Dr. Chakrabarty's resume on his website showed a $100,000 "gift" from Susman Godfrey L.L.P. The memorandum describes that entry as $100,000 in funding from Susman in December 2025, an "Unrestricted Gift for sponsored research" on "How AI generated books dilutes the market for human authors?"

They say that on Sunday, August 2, 2026, the day before expert Michael Lasinski's deposition and after expert-report deadlines, Susman served a supplemental report by Lasinski devoted to the paper. The memorandum calls Lasinski the class plaintiffs' expert on statutory damages and market harm under the fourth fair-use factor, 17 U.S.C. § 107(4). It says class plaintiffs then cited the paper in summary-judgment papers: as an exhibit, in 13 paragraphs of their statement of undisputed facts (paragraphs 858 through 871), and in their brief as an academic study, without disclosing the claimed funding.

The memorandum also says that on August 27, 2026, the public resume was changed and the Susman line removed. It quotes the revised resume as saying an earlier claim of a $100,000 unrestricted gift from Susman Godfrey LLP "was incorrect as the research was done for In re Mosaic LLM litigation for which [his] institution was compensated in a lesser amount."

What the class plaintiffs said

On October 2, 2026, the class plaintiffs filed a notice of nonopposition, signed by Justin A. Nelson of Susman Godfrey, interim lead counsel.

They wrote that the motion's premise, that counsel funded the paper, "is incorrect." Declarations from the two principal authors, they wrote, confirm the funding did not come from Susman Godfrey or any other class counsel. Susman retained Dr. Dhillon as an expert in this action and in In re Mosaic LLM Litigation, No. 3:24-cv-01451-CRB (N.D. Cal.), and retained Dr. Chakrabarty for Mosaic work, paying his institution for that work. A footnote says he was disclosed under the protective order in both cases but retained only in Mosaic, and that his institution received no compensation for work in this action.

The notice states: "Susman Godfrey did not fund the Generative AI paper, nor did it provide edits or comments to the paper." The class plaintiffs call the paper "at most, cumulative evidence." They point to other material they say is already in the record, including a statement by OpenAI's then-Policy Director about systems that "substitute for the labor of [] people," including "fiction authors . . . on Amazon," and Microsoft internal predictions of a "doom loop" from a product that "threatens the economic foundations of its essential suppliers."

They wrote that they "will withdraw their reliance on the Generative AI Paper and do not oppose the requested relief," citing the companies' concern about a report served after the rebuttal deadline and their view that the paper is cumulative.

What the order does not decide

Judge Stein granted the motion. The order does not decide who paid for the research. It does not resolve the summary-judgment motions. It does not remove the other market-harm material the class plaintiffs said is already in the record. It also does not take the paper off the internet. The companies told the court the authors had posted it, still labeled a "Working Paper Under Review."

Why it matters for authors

The ruling pulls one paper out of the class case. Class plaintiffs may not cite or rely on "Generative AI Floods and Dilutes the Market for Books," and the Lasinski supplement on that paper is struck. The order does not decide whether AI dilutes the book market, and it does not decide fair use.

The class lawyers said other market-harm evidence remains. On October 2 they called the paper cumulative and pointed to material already in the record, including the policy-director statement and the "doom loop" documents. This order does not strike those items.

A working paper is not a court finding. The companies said their motion was not a Daubert challenge. The order does not hold that the paper's conclusions are true or false, and it does not decide who funded the research.

See unsealed filings in the authors' OpenAI class action and the Third Circuit decision on Westlaw headnotes.

Sources

• U.S. District Court for the Southern District of New York: Order, October 6, 2026 (Document 2090)
• OpenAI and Microsoft: Notice of Motion, September 23, 2026 (Document 2045)
• OpenAI and Microsoft: Memorandum of Law, September 23, 2026 (Document 2046)
• Class Plaintiffs: Notice of Nonopposition, October 2, 2026 (Document 2087)