IP Law Daily, COPYRIGHT—S.D.N.Y.: Case over BloombergGPT passes a preliminary hurdle, (Dec 1, 2025)
Law Firms Mentioned:DiCello Levitt LLP | Fried, Frank, Harris, Shriver & Jacobson LLP
Organizations Mentioned:Bloomberg Finance, L.P. | Bloomberg L.P. | Fried, Frank, Harris, Shriver & Jacobson, LLP
By Matthew Hersh, J.D.
The court refuses to dismiss the claim on fair use grounds—at least for now.
A former presidential candidate and several other faith-based authors adequately stated a copyright infringement claim against the developer of a finance-based artificial intelligence large language mode, the federal court for Manhattan has held. The court, in dismissing a motion to dismiss the class action lawsuit, also found that it was too early in the case to determine whether the development of the financial chat program was protected by the doctrine of fair use (Huckabee v. Bloomberg, L.P., No. 1:23-cv-09152-MMG (S.D.N.Y. Nov. 24, 2025)).
The order comes in a class action lawsuit filed by a group of authors led by former governor, presidential candidate, and book author Mike Huckabee. The named plaintiffs assert a single copyright action alleging that Bloomberg L.P. and subsidiary Bloomberg Finance improperly used their copyrighted materials—including Huckabee’s best-seller “God, Guns, Grits, and Gravy”—to train and develop an artificial intelligence large language model known as BloombergGPT. In addition to Huckabee, the named plaintiffs include the authors of several Christian faith-based works, including Texas-based authors David Kinnaman and Tsh Oxenreider, Florida-based author Lysa TerKeurst, and Arkansas-based author John Blasé.
Bloomberg filed a motion to dismiss the complaint, leading to this opinion.
Statement of claim. The court found that the complaint, as amended, adequately stated a claim of copyright infringement. The class plaintiffs alleged that Bloomberg infringed their copyrights by importing into its database, and then using for training, all of the works within a pirate book database known as “Books3.” But although the class plaintiffs identified the books that they had each individually authored, Bloomberg pointed out, they did not identify which of those books had actually been contained within Books3 at the time Bloomberg accessed it. Nor did the complaint contain other details, Bloomberg argued, that would be necessary to state a claim. But the court rejected Bloomberg’s arguments.
The class plaintiffs adequately stated a claim, the court found, because their complaint put Bloomberg on fair notice of the allegations against it. Bloomberg argued that the class plaintiffs failed to allege “by what acts during what time” Bloomberg committed the alleged infringement. But the court had little trouble rejecting this argument. Under the federal rules, the court noted, a plaintiff is required to give only “a short and plain statement of the claim showing that the pleader is entitled to relief.” This rule “does not require every factual detail of every allegation,” the court noted. “Nor does it require a court to ignore common sense.” Here, the court emphasized, the complaint “(1) alleges that Books3 contains Plaintiffs’ copyrighted works; (2) cites a paper, released by Bloomberg, that stated Bloomberg used the data in Books3 to train BloombergGPT; and (3) states that this use occurred without Plaintiffs’ authorization or permission.” This was sufficient, the court reasoned, to state a complaint.
Nor did it matter that the class plaintiffs never specifically identified which of their copyrighted works were in Books3, the court found. The class plaintiffs, the court noted, alleged they hold copyrights in several specifically named works; that the Books3 data set includes their copyrighted works; and that Bloomberg used the materials in Books3 to train BloombergGPT without permission. “The clear and logical inference from Plaintiffs’ allegations, viewed as a whole, is that one or more of the specific copyrighted works attributed to each named Plaintiff in the Amended Complaint were included in Books3,” the court noted. “That is sufficient to survive a motion to dismiss.”
Fair use. The court also refused to dismiss the claim on grounds of fair use—at least at this point in the proceedings. While it is possible to decide a fair use defense on the face of a complaint, the court noted, courts more frequently address the defense at summary judgment, as “a defendant’s entitlement to a fair use defense must be clearly established on the face of a complaint to support dismissal.” Bloomberg’s claimed entitlement to a fair use defense did not meet this high standard, the court found, “because the Court needs a robust factual record to conduct the fair use analysis in this case.”
The court illustrated its point with an example using the fourth fair use factor—“the effect of the use upon the potential market for or value of the copyrighted work.” In order to analyze this factor, the court noted, it would conceivably have to address the following questions: “Will widespread copying of literary works to train LLM AI programs result in a deluge of AI-produced literary works, adversely affecting the market for the Works? Because LLM AI programs have the potential to yield massive profits but require significant amounts of text to train, is there a derivative market for literary works used to train AI? And, if so, does unlicensed copying of literary works detrimentally harm an author’s ability to participate in and profit from that derivative market?” These few examples, the court noted, “illustrate the futility of conducting a fair use analysis on the face of a complaint in this complex and evolving area of both law and technology, without a factual record.” Accordingly, the motion to dismiss would be denied.
The Case is No. 1:23-cv-09152-MMG.
Judge: Garnett, M.
Attorneys: Adam J. Levitt (DiCello Levitt LLP) for Mike Huckabee. Nicole M. Jantzi (Fried, Frank, Harris, Shriver & Jacobson LLP) for Bloomberg L.P. and Bloomberg Finance, L.P.
Companies: Bloomberg L.P.; Bloomberg Finance, L.P.
Cases: AINews Copyright TechnologyInternet NewYorkNews