Anthropic has responded to music publishers' demands for an injunction over AI training data, asserting that using copyrighted lyrics for AI training falls under fair use and has not caused "irreparable harm" to publishers.

The dispute centers on Concord Music Group and other publishers' allegations that Anthropic infringed copyrights by using their lyrics to train its Claude large language model (LLM).

Key points from Anthropic's response:

• Publishers haven't demonstrated "irreparable harm," which is required for an injunction
• Any potential damages can be addressed through monetary compensation
• The use of copyrighted works for AI training constitutes fair use through transformation
• An injunction would significantly impair AI model development
• The public has a vested interest in allowing AI innovation to continue

Anthropic has positioned itself as an ethical AI company by publicly releasing its system prompts for Claude LLM. Alex Albert, head of developer relations, indicated the company will continue sharing such information as system prompts are updated.

This legal battle extends beyond music publishers, as authors have also filed a class action lawsuit against Anthropic over the alleged misuse of their work in AI training.

Related Articles

Previous Articles