Music publishers Universal Music, ABKCO, and Concord have intensified their legal battle against AI company Anthropic, responding to the company's opposition to their preliminary injunction request in their copyright infringement lawsuit.

The publishers argue that Anthropic's Claude chatbot illegally trained on and reproduced protected song lyrics without authorization. Key points from their latest filing include:

1. Inadequate Safeguards

• Anthropic's new "guardrails" to prevent lyric reproduction are inconsistent and potentially circumventable
• These measures aren't a complete solution and could be removed at any time
• Publishers continue to find verbatim copies and unauthorized derivatives of protected lyrics

2. Responsibility for Infringement

• Publishers reject Anthropic's claim that they're responsible for infringement by submitting test queries
• These queries were necessary to detect copyright violations
• The prompts align with expected normal user behavior

3. Fair Use Challenge

• Publishers contest Anthropic's position that training AI models on copyrighted materials constitutes fair use
• They argue the use is commercial rather than transformative
• The practice threatens existing licensing markets and artistic control

4. Specific Concerns

• AI-generated derivatives and mashups pose unique challenges to potential licensing solutions
• Publishers highlight examples of error-filled or offensive outputs they "would not license at any cost"
• The system threatens to displace human songwriters with AI models built on their creativity

The publishers seek to make Anthropic's current safeguards permanent during litigation and prevent the company from using protected lyrics in future training models.

This case continues alongside other significant AI copyright disputes, including the recent partial dismissal of Sarah Silverman's class-action suit against OpenAI.

Related Articles

Previous Articles