Anthropic achieves Victoria Key to use just in demand for the author of AI

By Canuto

A historical decision from the Court of the Northern District of California gives Anthropic a key victory over the right use in artificial intelligence training, allowing the use of protected books – except pirate copies – for LLMS. However, the company must still face a damage to the use of pirate works in its central library. The ruling could transform the panorama of the development of AI and the treatment of copyright throughout the industry.
***

• The California Court ruled in favor of Anthropic regarding the use of protected books to train AI, considering this use as ‘fair’, except for pirated copies.
• Anthropic will face a judgment to determine compensation for damage associated with the use of millions of copies obtained illicitly.
• The decision could lay a precedent for the use of copyright material by AI platforms, generating impact on the entire technological industry.

---

Context: The legal battle between authors and artificial intelligence platforms

A high -profile legal conflict develops around the training of artificial intelligence models (AI) with copyright protected material. In August 2024, the authors Andrea Bartz, Charles Graeber and Kirk Wallace Johnson filed a collective claim in the Northern District of California against Anthropic, the company behind Claude AI, one of the most promising and backed by giants like Amazon and Google.

The lawsuit claimed that Anthropic had copied millions of books, both from pirate and bought sources, to integrate them into their huge digital library. The objective: Use these protected texts as input to train large language models (LLMS), without having the authorization of the authors.

In response, Anthropic argued in April 2025 that the cause was not adequate to group as collective demand, pointing out the complexity and diversity of the supposed damage to more than five million books. The company also warned about the financial ruin to face up to USD $ 150,000 in statutory damage for each of the books, which could add an astronomical amount of USD $ 750,000 million.

The failure: Train with protected works is fair use?

On June 23, 2025, the Court issued a transcendental decision. According to AI Fray’s original report, the Northern District of California granted a summary judgment in favor of Anthropic, stating that both the use of protected books to train AI, and the process of converting printed copies to digital format, constituted “fair use” in accordance with section 107 of the copyright law.

Judge William H. Alsup based his analysis on the four classic factors of fair use: the purpose of use (if it is commercial or educational), the nature of the work used, the amount used and the impact on the potential market or value of the work. The decision considered that, except for the second factor (the nature of the works, mostly literary), the other elements favored the anthropic position.

The court underlined the “transformative” character of AI technology, qualifying it as one of the most disruptive of the last decades. He also acknowledged that the digitalization of a purchased copy, if it implies destroying the original printed and not redistributing the digital file, can be considered a fair use under certain circumstances.

The exception: pirate copies lead Anthropic to Judgment for Damage

Despite the resonance that generates this partial triumph for Anthropic and for AI platforms in general, it was not an absolute victory. The Court determined that the pirated copies that the company compiled to be part of its central library could not be covered in the figure of fair use, especially when retaining those “forever” files and “general purposes”, even if not all were used in LLMS training.

At this point, the ruling was overwhelming: “Each factor points against fair use” regarding pirate works, highlighting that no argument justified that use beyond economic convenience for Anthropic. The Court also clarified that the subsequent purchase of legitimate copies of a book does not exculpate the company of the previous illicit use, although it may affect the amount of the damage determined in the following phase of the process.

Consequently, a trial was scheduled to define the responsibility of Anthropic and the amount of damages related to the use of these pirated copies. Given the potential magnitude of the figures involved, the final decision could drastically impact the company’s finances and mark a before and then in the industry.

Repercussions for the AI ​​industry and future legal implications

Although the current order comes from a district court and is limited to a specific case, its resonance transcends the entire technological industry. If the interpretation is maintained, the door would open so that other companies use protected material to train their systems, as long as they do not resort to pirate sources and meet certain requirements such as transforming and not redistributing the content.

The decision has been classified by experts as mixed: a partial loss for both rights holders and Anthropic herself, but also a possible legal framework for the evolution of generative AI. The implications are not only concentrated in this process, but also in several similar litigation – as the case of Authors Guild against Openai in New York and the action of Thomson Reuters against Ross Intelligence in the Court of Appeals of the Third Circuit – where the delicate balance between the technological advance and the defense of intellectual rights is discussed.

It is important to note that the authors could still appeal the ruling on fair use, although most likely this happens only after the judgment on damage and final sentence. On the other hand, Anthropic’s position in front of the court will be carefully observed throughout the industry, being one of the best financed startups and one of the pioneers in the application of large language models globally.

Conclusion: What does the failure for the future of AI mean?

This case represents a signal of possible flexibility in the doctrine of fair use for AI, but leaves a clear warning about the use of illegally acquired works. The world of artificial intelligence and copyright remains, therefore, in a gray area, in which each new decision can redefine the limits between innovation and intellectual property.

Waiting for the determination of damages in the coming trial, the eyes of the technological, legal and creative sector will continue in California and in the next decisions that require the scope of this preliminary verdict.

---

Original image of Diariobitcoin, created with artificial intelligence, for free use, licensed under public domain.
This article was written by an AI content editor and reviewed by a human editor to guarantee quality and precision.

WARNING: Diariobitcoin offers informative and educational content on various topics, including cryptocurrencies, AI, technology and regulations. We do not provide financial advice. Cryptactive investments are high risk and may not be adequate for all. Investigate, consult an expert and verify the applicable legislation before investing. I could lose all its capital.