↓Skip to main content
Zeitgeist — a spike by Chris Gathercole
  1. Topics/
  2. Data, IP & Training Rights/

Gather — 9 May 2026

Publishers vs Meta — Mainstream Press Coverage #

  • Publishers sue Meta, claiming it violated copyrights in training AI with their books (Washington Post, 2026-05-05) — WashPost’s coverage of the Elsevier/Cengage/Hachette/Macmillan/McGraw Hill + Scott Turow suit against Meta. Notably emphasises that Llama is open-weight: if open-weight models carry training data liability, redistribution becomes a liability vector for every downstream user and fine-tuner, not just Meta — a structural difference from closed-model suits.
  • Scott Turow, Macmillan, McGraw Hill sue Meta for AI copyright infringement (NPR, 2026-05-05) — NPR’s angle: Scott Turow as the named public-facing plaintiff is a strategic choice by the coalition — a recognisable author (and Authors Guild president) attached to what is otherwise a corporate publisher lawsuit. The same Turow who brokered the Bartz/Anthropic settlement is now leading a parallel suit.

Bartz Final Approval — Imminent Checkpoint #

The $1.5B Bartz v. Anthropic settlement goes to final approval on May 14, 2026 at 2:00 p.m. PT. If the court formally endorses the dual holding — training = fair use; piracy = not fair use — the $3K-per-work reference price becomes explicitly precedential and will be cited in every subsequent training-data negotiation. The May 14 ruling is the most consequential near-term milestone in AI copyright law.

Music Licensing — The Divergent Track #

  • Licensed or lost? In the future of AI training, “the world is splintering” (Music Ally, 2025-12-08) — The music industry is negotiating licensing frameworks rather than suing for training use — usage-based royalties, licensed catalog access, and consortium structures. This is a fundamentally different negotiating stance from publishers, whose default is litigation. Music Ally’s “world is splintering” framing: music, books, news, and academic publishers are each developing distinct IP responses to AI training with no converging framework in sight.
  • Open vs Closed AI Ecosystems WashPost’s emphasis on Llama being open-weight is the key cross-link: closed models have a single liable entity; open-weight models distribute liability to every redistributor and fine-tuner. If the Meta suit succeeds, open-weight model distributions could carry attached training data liability.
  • AI Impact on Society Scott Turow leads both the Bartz settlement (as Authors Guild president) and the Meta suit — the same organisation operating as settlement broker and litigation plaintiff in parallel, a dual-track strategy that signals the Authors Guild views both settlement and litigation as complementary levers.

Meta-observations #

  • Emerging pattern: Mainstream press (WashPost, NPR) now covering individual publisher AI lawsuits as public-interest stories with named authors as protagonists — no longer confined to legal press. The frame has shifted from “big tech vs copyright” to “specific books, specific harm,” which is more sympathetic to plaintiffs.
  • Gap: Music industry licensing track remains structurally undertracked. Music Ally (Dec 2025) is the best available framing; adding a dedicated keyword would catch the licensing-deal track that is developing in parallel to the litigation track.
  • Keyword suggestion: "AI music licensing" deals OR royalties 2026 — fills the music track gap flagged in 2026-05-06.