The Digester

Encyclopedia Britannica sues OpenAI for using its articles to train ChatGPT

Mar 16th 2026

Britannica and Merriam-Webster say OpenAI copied thousands of reference entries to train ChatGPT and seek damages and an injunction, while OpenAI says its training relies on publicly available material and fair use.

  • Britannica and its Merriam-Webster subsidiary filed suit in Manhattan federal court alleging OpenAI copied nearly 100,000 articles to train its GPT models.
  • The complaint says ChatGPT can produce near-verbatim encyclopedia entries and has diverted web traffic with AI-generated summaries.
  • Britannica also accuses OpenAI of trademark infringement and of falsely citing Britannica in AI hallucinations.
  • OpenAI said its models are trained on publicly available data and are grounded in fair use.

Articles