Archive — US — turtlesai
AI and Books: Anthropic Wins on Fair Use but Takes a Big Risk for Piracy
Anthropic receives a favorable ruling on fair use for training AI with purchased books, while facing a separate trial for illegally using over 7 million pirated texts. The court emphasizes that, although transformative use is permitted, piracy is impermissible and carries potential damages, setting a precedent for AI and copyright law.
Key points
- Judge Alsup rules that digitizing purchased books is fair use as it's transformative.
- Anthropic's practice of using pirated books is not justified and will face legal scrutiny.
- A separate trial in December will determine damages for the 7 million pirated copies.
- Anthropic's initial strategy was to negotiate licenses but opted for cheaper pirated content.
- This ruling impacts other AI companies regarding copyright and usage of training data.
Why it matters
- This case underscores the ongoing debates over copyright law in the digital age.
- Big tech companies face similar challenges regarding intellectual property and AI.
- The ruling aligns with trends towards stricter copyright enforcement amid rising AI development.
- The case reflects a broader shift in legal interpretations of fair use in AI-model training.
The ruling sets a crucial precedent for the intersection of AI and copyright law, allowing fair use for purchased texts but highlighting the significant risks associated with using pirated materials, which will impact how future AI training practices are structured.