A Fair Use Fault Line: What the Anthropic Ruling Means for IP Valuation in the Age of AI

A recent decision from the Northern District of California in Bartz et al. v. Anthropic PBC provides new guidance on how U.S. courts may handle copyright and fair use disputes involving AI training data. For companies and professionals involved in intellectual property valuation, this ruling clarifies the boundary between lawful use and infringement, especially as generative AI systems increasingly rely on copyrighted materials.

Case Overview

The plaintiffs, a group of published authors, sued Anthropic, the AI firm behind Claude, for copyright infringement. Anthropic admitted to:

• Downloading millions of books from pirate sites

• Scanning purchased physical books and retaining them for a central digital library

These works were then used to train large language models (LLMs). The court was asked to determine whether these uses qualified as fair use. The ruling depended on the type of use.

Key Findings

The court found that training LLMs like Claude was “exceedingly transformative”, qualifying as fair use under Section 107. Claude did not output infringing material—no excerpts or copied texts were shown to users.

Digitizing legally purchased books for internal research was also deemed fair use. The conversion from print to searchable digital files was considered non-exploitative and transformative, offering benefits such as storage efficiency and search functionality.

In contrast, the court rejected Anthropic’s use of pirated books. Retaining unauthorized copies was not transformative and was deemed a cost-avoidance strategy with no legal protection.

Implications for IP Valuation

From a valuation perspective, this ruling reinforces several important points:

• Transformative Use May Reduce Liability: Using copyrighted materials in ways that alter the work’s purpose, such as training AI instead of distributing content, may be viewed favorably by courts.

• Provenance and Licensing Matter: Firms relying on unauthorized content—even for internal purposes—face legal, reputational, and financial risks.

• Ongoing Legal Uncertainty: Questions remain about derivative works, other types of IP, and secondary uses, emphasizing the need for continued legal monitoring.

Conclusion

The Anthropic decision is not the final word on copyright and AI, but it sets a precedent for IP strategy, risk assessment, and valuation. Companies developing or investing in AI technologies must ensure proper sourcing of training data and document fair use positions carefully. These steps are essential for both legal compliance and protecting enterprise value in M&A, licensing, and funding transactions.