TRI Thomson-Reuters Corp

$TRI Only 920 followers. 400,000 firms. 1,000,000 CoCounsel users. 95% of Fortune 100 and 97% of the 100 top grossing firms in the US. Decades of proprietary structured legal data saving lawyers and accountants time (and employees) via Anthropic models. Thomson-Reuters doesn't sound like an data/agentic AI play, but it is. It dumped half earlier this year because of AI fears. Well, yes, AI fears... but who has the data? $TRI has the data. Turnaround beginning?

$TRI data point: a Michigan class action alleges Thomson Reuters' CLEAR and Westlaw tools showed five digits of users' Social Security numbers, past the four allowed by state law. Complaint stage, nothing to claim. Open Class Actions.

$TRI The first US appeals-court ruling on AI training and copyright landed Tuesday: the 3rd Circuit upheld Thomson Reuters' copyright win against Ross Intelligence and rejected the fair-use defense. The case goes back to 2020. Ross allegedly copied thousands of Westlaw headnotes, the short summaries of points of law, to train an AI legal search engine. Ross shut down in 2021, citing the litigation cost, so the ruling is now mostly about precedent. Two caveats. First, this was not a generative AI case. Ross built a retrieval engine, and the tech companies in the NYT and music-label lawsuits are leaning on fair use precisely because their systems transform training data into something new. Second, the court kept its reasoning sealed, so other courts get limited guidance from this. Shares fell 1.1% Tuesday. The DOJ sided with OpenAI on fair use in the NYT case this month. First appeals-level datapoint, but not the decision that settles the generative AI cases.

$TRI They have the data. They have the relationships (400,000 firms... hello?). Looks like a treasure to me, like $NICE