Anthropic Engaged Datamation to Destructively Scan Millions of Books for Claude Model Training
What happened
Court records confirm Anthropic's 'Project Panama' involved destructively scanning millions of purchased print books for Claude model training, intensifying the debate over fair use and intellectual property in AI data acquisition. This effort, internally described as 'destructively digitizing' books, highlights the aggressive data collection practices of major AI labs.
Why it matters
Policymakers must recognize that current fair use interpretations may permit destructive digitization of purchased works, but the substantial $1.5 billion settlement for piracy underscores escalating legal and financial exposure for AI companies.
Topics
- AI Model Training
- Destructive Book Scanning
- Copyright Fair Use
- Data Provenance
Articles in this trend
- Claude: Anthropic engaged Datamation to strip, cut, and scan purchased print books into digital form for model training. — Pascal’s Substack
- The Danger of “Just Scrape It” in AI Strategy — HackerNoon
- Microsoft’s Nadella calls out Big AI for hypocrisy — but what about his own company? — Computerworld
- OpenAI wins first battle in ongoing Indian copyright lawsuit — Technollama
- The Problem of Provenance: Nobody Knows What’s In the Training Data — Machine Learning on Medium
- Who’s Afraid of Chinese Models? — Simon Willison's Weblog
- When the AI bubble bursts, what will Australia do with the tools it built? One man thinks he has the answer — AI (artificial intelligence) | The Guardian
- Learn the Ethics of AI — AI on Medium