What happened

  • The case was filed in July 2023 and consolidated under Master File No. 5:23-cv-03440-EKL. Google's brief says the deadline to add parties passed on December 20, 2024, and that in September 2025 the court closed the pleadings, finding further amendment "would be unduly burdensome on Defendants and would further delay resolution" (Dkt. 505, quoting Dkt. 216).
  • Plaintiffs moved for class certification in October 2025 for a Books Class and an Images Class of legal or beneficial owners of registered copyrights (Dkt. 445). Google opposed (Dkt. 446). Briefing finished at the end of 2025.
  • On January 15, 2026 two publishers, Cengage Learning and Hachette Book Group, moved to intervene to seek a publisher subclass (Dkt. 342). They later withdrew and filed their own putative class action against Google in the Southern District of New York (Hachette Book Group v. Google, No. 1:26-cv-05870). The court granted the withdrawal and wrote that the publishers' concern about authors' adequacy "was a deficiency of the publishers' own invention" (Dkt. 477).
  • On February 18, 2026 the court stayed the unexpired case deadlines and discovery pending its ruling on class certification and related motions, which it heard on February 20 (Dkt. 421). Fact discovery had closed on February 13, 2026, according to Google.
  • On September 28, 2026 seven textbook authors, six individuals and one company, moved to intervene (Dkt. 490) to file a complaint adding new allegations of torrent-based distribution and 84 more works, and to lead a textbook-author subclass. Plaintiffs do not oppose their filing a complaint but oppose adding them to class leadership. Google opposes intervention as untimely. The hearing is set for January 6, 2027 (Dkt. 504, 505).
  • On September 29, 2026 the magistrate judge denied without prejudice a February 2026 privilege-log dispute, noting that the action "remains stayed" (Dkt. 499).

What the court said

The court has not ruled on whether a class can be certified, so there is no holding on whether Google's training was fair use or infringement. What the orders show is procedural: the court stayed everything because "the outcome of these motions will impact further proceedings in this case" (Dkt. 421), and it treated the publishers' attempt to take over class leadership skeptically, noting that Google, "the only party opposed to class certification," had not questioned the authors' adequacy (Dkt. 477).

Google's central argument against certification, as its October filing summarizes it, is that ownership, registration, licenses, limitations, fair use and damages would require work-by-work proof. That is Google's position, not a finding.

What it means for your firm

  • Our reading: the class certification ruling is the next real signal in this case. If a class is certified, copyright claims over AI training become a single large claim; if not, the claims fragment into individual suits and separate actions like the publishers' New York case.
  • Work-by-work proof is the battleground. Rights holders advising authors, artists or publishers need clean records of who owns and registered each work. Clients using Google's models need to know whether a class could reach them.
  • Parallel filings multiply. The publishers sued separately in New York and textbook authors are threatening a separate suit, so one vendor can face several related cases in different courts.
  • Nothing here decides fair use. Do not cite this docket as authority on training. Cite the rulings that decided it, such as Bartz v. Anthropic and Kadrey v. Meta, and check them for their limits.

What to do Monday

  1. If you represent an author, artist or publisher, confirm for each work that you hold registration details and a clear owner record, since class membership here depends on registered copyrights.
  2. If you advise a client that uses Google's generative AI tools, ask for the vendor's current indemnity and notice terms for copyright claims about training data.
  3. Add the Northern District of California and Southern District of New York dockets for Google to your watch list, and note the January 6, 2027 intervention hearing.
  4. Before relying on this brief, pull the live docket for No. 5:23-cv-03440-EKL and check for a class certification order entered after October 19, 2026.

Sources

Questions and answers

What is the Google generative AI copyright case about?

Authors and artists allege Google used their copyrighted books and images without permission to train its generative AI models, and seek to represent classes of owners. It was filed in July 2023 in the Northern District of California.

Has the court decided whether Google infringed copyright?

No. In the filings we read, which run to October 19, 2026, the court had heard argument on class certification on February 20, 2026 and had not ruled. The case is stayed.

Why are publishers and textbook authors trying to join?

Cengage and Hachette sought to lead a publisher subclass, then withdrew and sued separately in New York. Seven textbook authors asked on September 28, 2026 to intervene and lead a textbook-author subclass, arguing the current plaintiffs do not represent them. Google says the request is too late.

Operational information, not legal advice. Facts come from the filings on the court docket linked under Sources, read on Oct 27, 2026. A docket moves: this brief covers only the filings we read and does not replace the live docket. The reading of what it means for your firm is ours.