Anthropic has submitted a 40-page response opposing the preliminary injunction requested by Universal Music, Concord, and ABKCO in their ongoing copyright infringement lawsuit. The dispute centers on Anthropic's use of copyrighted music compositions during AI training and their AI assistant Claude's ability to reproduce song lyrics.

Key points from Anthropic's response:

• The company maintains that using copyrighted content to train large language models (LLMs) qualifies as fair use
• Anthropic argues there's no evidence of irreparable harm to publishers, noting they waited months before taking legal action
• The company attempted to implement additional guardrails and cooperate with publishers before the injunction filing
• Anthropic contends that requiring licenses for training AI on copyrighted content would make general-purpose AI tools impossible to develop, citing the massive scale of data needed

Technical and Legal Arguments:

• Anthropic challenges the venue, stating they have no relevant connection to Tennessee
• The company argues that any alleged infringement was actually caused by the plaintiffs' own prompts to Claude
• They assert that monetary damages would be sufficient if publishers ultimately prevail in the case

The case highlights the growing tension between AI companies and content creators over the use of copyrighted materials in AI training. Similar suits are currently pending across the country, though none have resulted in preliminary injunctions.

Anthropic's response emphasizes that song lyrics are not among typical user requests for Claude, and the company has implemented additional safeguards to prevent unauthorized lyric display in response to the controversy.

Related Articles

Previous Articles