Authors Guild v. OpenAI September 2023: When the Major Book-Author Class Action Joined the Generative-AI Training-Data Copyright Docket

Incident summary

On 19 September 2023 the Authors Guild and seventeen named author plaintiffs filed a class-action lawsuit against OpenAI in the United States District Court for the Southern District of New York (Case No. 1:23-cv-08292). The named plaintiffs included Jonathan Franzen, John Grisham, George R.R. Martin, Jodi Picoult, Scott Turow, and David Baldacci. The complaint alleged that OpenAI trained GPT models on the plaintiffs' copyrighted books, with the training data including book corpora like Books3 that contained substantial swaths of contemporary fiction without license.

The Authors Guild filing joined a growing docket of generative-AI training-data copyright litigation. The Sarah Silverman, Christopher Golden, and Richard Kadrey class action against OpenAI had been filed 7 July 2023 in the Northern District of California (Case No. 3:23-cv-03416). The New York Times v. OpenAI suit followed on 27 December 2023.

Through 2024 and 2025 the cases proceeded through motions to dismiss, narrowing of claims, and discovery. The Authors Guild case is one of several actions that have been consolidated in some venues. The legal questions (whether training constitutes fair use, whether memorization and reproduction of training material is infringement, whether RAG-style retrieval is distinct from training) are foundational and unresolved at the time of writing.

Failure technique

The legal-technical pattern is unauthorized reuse of copyrighted text in foundation-model training where the trained model can reproduce substantial passages from the training material. The Books3 dataset (compiled from a controversial pirated-books collection) was a primary evidence target: plaintiffs alleged Books3 was used in OpenAI training and that GPT models could be prompted to reproduce passages from books in Books3.

Per the US Copyright Office July 2024 report on Generative AI Training, the central open questions are whether training is a transformative fair use, whether memorization changes the analysis, and whether commercial vs. non-commercial training context matters. The Authors Guild and parallel cases are foundational litigation that will establish US copyright doctrine for AI training.

OpenAI's defense has consistently asserted fair use, transformative purpose, and that the trained model does not reproduce specific copyrighted material at scale. Discovery and motion practice in the consolidated cases have tested both factual and legal aspects of these defenses. The litigation is being watched as the foundational text-training-data copyright case in parallel with Getty v. Stability for image-training data.

Impact and consequences

Direct commercial impact on OpenAI during 2023-2025 has been substantial in legal cost and operational distraction, though the company has continued to grow during the litigation period. OpenAI has executed direct content licensing deals with major publishers (Associated Press, Axel Springer, News Corp, Time, Vox Media, others) during the litigation period; these deals are partially a market response to the litigation risk.

Industry impact: the Authors Guild case is one of the foundational text-training-data copyright cases. Subsequent foundation-model launches (Anthropic Claude 3, Google Gemini, Meta Llama 3, others) have all addressed training-data sourcing more explicitly than 2022-era products did. Direct content licensing of major news and publisher corpora has become a competitive surface among foundation labs.

Legal-precedent impact: trial outcomes in the Authors Guild and parallel cases will be foundational for US AI copyright doctrine. The cases sit alongside the NYT v. OpenAI case (specific facts about article reproduction) and the Sarah Silverman case (book reproduction) to test multiple aspects of the training-data IP question.

Lessons for builders

Treat training-data IP review as a launch-readiness gate for commercial foundation-model products serving production workloads. The post-2023 commercial landscape has shifted toward direct content licensing (OpenAI publisher deals), explicit indemnification (OpenAI Copyright Shield, Microsoft AI Customer Copyright Commitment), or open-licensing-only training data. AI Strategy Lead and Senior AI Product Manager own this gate.

Watch for memorization on the model-output side. The cases consistently use evidence of model reproduction of training material as the bridge from corpus claim to model claim. Pre-deployment evaluation should test for memorization of identifiable training-corpus signatures (specific sentences, character names, distinctive passages).

Track the US Copyright Office and EU AI Act developments. The US Copyright Office's July 2024 Part 3 report on Generative AI Training is the most-cited US-policy artifact; the EU AI Act Article 53 training-data documentation obligations apply to general-purpose AI models. Both shape the legal-defense surface.

Document corpus-assembly decisions and the licensing posture for each major corpus component. The Books3 evidence in the Authors Guild case is concrete; equally specific documentation on the licensing side is the operational defense.

Mitigations

What builders should put in place to address the failure pattern. Each mitigation maps to operational practice the relevant Applied AI roles own.

Related Applied AI roles

The Applied AI roles whose day-to-day work would have prevented, detected, or contained this incident.

• AI Strategy Lead: An AI Strategy Lead owns organizational AI strategy and prioritization at the company level.
• Senior AI Product Manager: A Senior AI Product Manager owns AI product strategy across multiple feature areas.
• AI Product Manager: An AI Product Manager owns AI-powered product features and the roadmap that ships them.
• AI Engineer: An AI Engineer builds production cybersecurity-relevant AI systems integrating LLMs, embeddings, and retrieval pipelines.

Companies central to this incident

Read the DecipherU Applied AI company profiles for the organizations whose decisions, products, or research shaped this incident.

• OpenAI: Frontier large language models and consumer + API AI products

Related AI Decipher Files

• New York Times v. OpenAI (Dec 2023): The Copyright Case That Defines AI Training Liability
• Stability AI v. Getty Images February 2023: When Image Generators Faced Their First Major Training-Data Copyright Lawsuit
• Adobe Firefly Training Data Controversy April 2024: When 'Ethically Trained' Claims Met Disclosed Use of AI-Generated Images

Frequently asked questions