Thomson Reuters v. ROSS Intelligence
Early AI-adjacent training decision cited in LLM cases for market harm and non-transformative commercial use.
Procedural posture
Bench trial and appeal on whether copying Westlaw headnotes to train a legal research tool was fair use; significant precedential weight (placeholder).
Facts
ROSS copied Westlaw headnotes and key numbering to train a competing AI legal research product (placeholder — verify findings).
Legal issues
- Intermediate copying to build a non-expressive training set
- Transformative purpose vs. commercial substitution
- Market harm to Thomson Reuters' legal research products
Holding
District court found infringement; fair use rejected on market substitution grounds (placeholder — verify post-trial order).
Reasoning
Court emphasized ROSS sought to market a competing research product using Thomson Reuters' expressive headnotes, undercutting transformative use arguments (placeholder).
Why it matters for AI law
Early AI-adjacent training decision cited in LLM cases for market harm and non-transformative commercial use.