Federal Court Finalized Anthropic Settlement
A July 2026 court ruling permits AI firms to destroy physical books after digitizing them for training models.
Updated on Oct. 4, 2026 in Artificial Intelligence
Live Poll
Should AI companies be permitted to destroy physical books to train their digital models?
In July 2026, a federal court finalized a class-action settlement involving AI developer Anthropic. The agreement permits companies to purchase books for digital scanning and subsequent destruction under the first-sale doctrine.
Why it matters
Tech giants utilize books as primary data sources to train large language models. The legal resolution formalizes the process for companies to acquire and process physical literature for digital use.
Companies are permitted to slice book spines to facilitate digital scanning, with requirements to recycle the physical remnants. The process involves bulk procurement of texts to feed data into large language models.
The players
Anthropic
Anthropic is an AI developer focused on building large language models and artificial intelligence safety.
Tom Congalton
At age 74, Tom Congalton is an experienced book collector who operates Between the Covers Rare Books in Gloucester Heights.
Antiquarian Booksellers' Association of America
This organization represents professional rare book dealers but has not declared a formal position regarding AI.
The details
The settlement allows AI companies to treat purchased physical books as digital inputs by destroying them after scanning. This practice impacts inventory management for businesses like Between the Covers Rare Books in Gloucester Heights, which catalogs over 250,000 volumes.
Timeline
Tom Congalton began collecting books in 1968.
John Grisham acknowledged the Congaltons in Camino Island in 2017.
A federal court finalized the class-action settlement involving Anthropic in July 2026.
The Tech Race
This settlement highlights a pivotal shift in the AI arms race where the acquisition of high-quality data becomes as critical as computing power. By utilizing the first-sale doctrine, companies are standardizing a process to bypass traditional licensing by physical destruction of assets.
This ruling establishes a legal pathway for AI developers to mass-digitize literature, potentially accelerating the capabilities of future AI models. Readers who own physical books should be aware that the legal landscape for copyrighted works in digital training is becoming more clearly defined.
The takeaway
The settlement creates a clear legal precedent for AI developers to process physical assets for digital intelligence training. This shift suggests that the physical format of knowledge is increasingly being commoditized as raw data for silicon-based systems.
Further reading
For broader insights into how industry standards are evolving, visit our Artificial Intelligence section.
Source note: This article includes information reported by New Jersey Monthly.
Live Poll
Should AI companies be permitted to destroy physical books to train their digital models?
