OpenAI's internal Slack messages could cost it billions in copyright suit
a day ago
- #legal-battle
- #AI-training
- #copyright-infringement
- OpenAI faces potential billions in penalties due to internal Slack messages discussing the deletion of a pirated books dataset.
- Authors and publishers suing OpenAI seek access to attorney communications, invoking a 'crime-fraud' exemption to privilege.
- Evidence includes OpenAI's use of LibGen, a pirated book database, mentioned in early research papers.
- Plaintiffs allege intentional destruction of evidence, which could lead to 'willful infringement' charges and penalties up to $150,000 per work.
- OpenAI denies waiving attorney-client privilege, stating decisions were made in consultation with counsel.
- Judge Ona Wang ordered some communications to be produced while others remain withheld.
- Meta's internal messages revealed similar concerns about using pirated datasets, approved by higher-ups.
- Anthropic settled for $1.5 billion over copyright claims, excluding pirated works from 'fair use'.