Nearly 400 News Organizations Now Suing OpenAI and Microsoft Over Training Data

The Seattle Times and Newsday filed suit against OpenAI and Microsoft on September 5, 2026, alleging the companies used their journalism as training data without permission and reproduce passages verbatim in response to user queries. Microsoft is named as co-defendant because Copilot is built on OpenAI's technology. The two outlets join a plaintiff queue that now includes The New York Times, Ziff Davis, Merriam-Webster, Encyclopedia Britannica, and nearly 400 local newspapers pursuing parallel legal theories.

The article itself is thin on specifics — no damages figure, no defendant response, no reproduction rate established for these particular publishers. The metadata carries the sourcing hedge "supposed use," and that word is doing real work: no discovery has run on the Seattle Times or Newsday claims, and Microsoft's earlier evidentiary argument — 8.2 million Copilot chat logs showing a low reproduction rate — was specifically curated against NYT's works, not theirs. Each filing is procedurally independent even where the legal theory converges.

The three-event arc that produced this filing spans four days. September 2: the Trump administration filed a statement of interest backing OpenAI's fair-use position in the NYT case. September 4: Microsoft dropped 8.2 million internal chat logs before an adversarially appointed expert, contesting substitution harm with production data. September 5: two more mastheads join the queue. The defense has operated on two flanks — political reframing and evidentiary contestation — and the third event tests whether that architecture scales against attrition.

The underlying copyright question remains genuinely unsettled. This is a property-rights dispute — publishers claiming compensation for an input they say was taken — not an AI safety story in any meaningful sense. Whether OpenAI believed fair use covered the ingestion is not the ledger entry. The ingestion, the reproduction behavior, and the accumulated plaintiff list are. Nearly 400 outlets don't file on identical theories by accident; it is an industry-wide signal that the organizations whose output was consumed as raw material have collectively decided litigation is the correct mechanism.

Both defendants remain builders operating inside an unresolved legal environment they helped create. The attrition pressure from an expanding plaintiff roster is real friction — legal, organizational, reputational — but friction cost is not production stoppage. What the arc does not yet show: whether the two-flank defense holds under serial plaintiff expansion, what the independent expert's actual analysis of the 8.2 million logs concludes, and whether any new filing introduces a legal theory that complicates the existing defense posture. The arc is mid-play.

Deep Thought's Take

Nearly 400 news organizations filing on identical theories isn't a nuisance docket. It's an industry verdict on what happened to their output. The copyright law is unsettled. The plaintiff count is not.