Skip to content
Submit incident
AII-1001-19419
Under investigationA regulator, court or police has opened a formal inquiry and reached no finding yet.

Meta Defends Use of Pirated Books to Train Llama AI Model

Meta faces ongoing litigation from multiple authors who claim the company used millions of pirated books downloaded from illegal libraries to train its Llama AI model. Meta has admitted to using the pirated content but argues this constitutes fair use, and further contends that automatic file redistribution through BitTorrent was an unavoidable technical requirement of the download protocol.

10 March 2026 · Meta · Intellectual property and appropriation
10 March 2026
AII-1001-194192 sources ↓
Quick facts
Category
Intellectual property and appropriation
Country
United States
Source
Economie Matin
Date
10 March 2026
Drafted by
Raidu Incident Detection

What happened

In 2023, authors Richard Kadrey, Sarah Silverman and Christopher Golden filed a class action lawsuit against Meta for using pirated books from illegal libraries as training data for its Llama language model.

Meta downloaded over 81.7 terabytes of data from pirate libraries including LibGen and Z-Library to train its AI models.

In March 2026, Meta argued in court that using BitTorrent to acquire and automatically redistribute pirated books constituted fair use, claiming that redistribution through BitTorrent was a technical necessity rather than a deliberate choice.

Reported impact

Authors alleged that Meta systematically obtained complete copies of protected works from pirate libraries without authorization or compensation to rights holders.

The plaintiffs contend that Meta wanted access to books specifically for their expressive content, which copyright law protects, but obtained this content through piracy instead of payment.

Index analysis

In summer 2025, a federal California court ruled that the training itself qualified as fair use. The remaining dispute concerns whether the redistribution of files through BitTorrent protocol, which automatically occurs when downloading via BitTorrent, also falls under fair use protection.

Meta asserted that the BitTorrent protocol mandates simultaneous file-sharing by all users who download, making redistribution an unavoidable technical consequence rather than deliberate infringement.

The authors' lawyers challenged Meta's defense as a procedural maneuver, noting the fair use argument regarding BitTorrent seeding had not been explicitly raised until after discovery deadlines, despite discussions in November 2024.

Governance controls

Meta claimed in its December 2025 case management filing that the technical necessity defense was disclosed within required procedures, disputing the plaintiffs' allegation of late disclosure.

Classification

Category
Intellectual property and appropriation
Industry
Media, social platforms and marketing
Country
United States
Harm status
Actual harm
Intent
Intentional
Timing
Unknown
Harm types
Economic or property
Harmed
Business
Failure modes
Training data misuse
Consequences
Litigation
AI systems
Llama
Regulations
FCRA (cited)

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Economie Matin
Meta assume le piratage de livres pour entraîner son IA
10 March 2026
Les Numériques
Meta estime que pirater des livres via BitTorrent pour entraîner son IA est tout à fait normal
10 March 2026
Jul 2026Nigeria opens investigation into tech giants over media content exploitationUnder investigationIntellectual property and appropriationYahoo actualitésJul 2026Delhi High Court Restrains AI and Deepfake Misuse of Ravi Kishan's IdentityConfirmedIntellectual property and appropriationThe Times of Indiathehindu.comAsian News International (ANI)+1Jun 2026Bombay High Court allows Preity Zinta to sue over AI-generated deepfakesConfirmedIntellectual property and appropriationThe Times of IndiaThe Indian ExpressTimesNow+5Jun 2026Judge allows copyright lawsuit over Meta's use of pirated adult films for AI trainingUnder investigationIntellectual property and appropriationMashableInternational Business Times UKMay 2026Journalists and narrators sue tech giants for using voiceprints to train AIAllegedIntellectual property and appropriationnews.bloomberglaw.comCourt House News ServiceBiometric Update+4