A federal judge in California granted preliminary approval on September 25 of a $1.5 billion settlement resolving a class action lawsuit by authors alleging Anthropic used their copyrighted work without permission to train generative AI models. This marks the first settlement in ongoing litigation against multiple AI vendors (OpenAI, Microsoft, Meta) over training-data provenance. The precedent establishes that AI vendors face quantifiable IP liability exposure and sets a legal standard for training-data claims that affects cost accounting and procurement risk for vendors unable to document clean data licensing.
318 other tracked developments touching the same claims.