Three publishers sue xAI — now SpaceX Super Intelligence — over Grok training on pirated and scanned books
04.11.2026 · Automatisch aus öffentlichen Quellen erfasst. Nicht von einer Person geprüft.
Dieser Beitrag liegt nur in seiner Originalsprache vor (Englisch).
What happened
On 4 November 2026, three copyright complaints were filed in the US District Court for the Northern District of California against the company formerly known as xAI, named in the filings as SpaceX Super Intelligence Corp., SpaceX Super Intelligence LLC and SpaceX Super Intelligence Holdings LLC, each listed as formerly X.AI. The plaintiffs are academic publisher Cognella (case 3:26-cv-13189), Chicken Soup for the Soul (3:26-cv-13188), and Cambronne with four individual authors (3:26-cv-13187). All three are brought by Stris & Maher with Freedman Normand Friedland and run the same three counts.
The counts
Count I alleges the company took the works from pirate libraries. Count II alleges it bought physical copies, scanned them, ran the images through optical character recognition, then "processed, cleaned, de-duplicated, tokenized, stored, and retained" the digital copies to train and evaluate the Grok family. Count III is contributory infringement. Each complaint demands a jury trial, and the plaintiffs argue that statutory damages must be set by a jury rather than the court.
Why Count II matters
The complaints argue that buying a book authorises none of the scanning: "Ownership of a physical copy conveys no right to reproduce the copyrighted work embodied in it." That is a direct challenge to the holding in Bartz v. Anthropic, where Judge William Alsup found that converting lawfully purchased print books into digital copies for training was fair use. The complaints cite that case for the facts of the scanning programme and ask a different court to come out the other way.
Evidence cited
The complaints lean largely on the defendant's own material, including its EU AI Act disclosure for Grok 4.7, published in September 2026, which states the model was pre-trained on more than 10 trillion tokens, that its crawler drew on "community and document-sharing sites", and that the company uses the US Trade Representative's Notorious Markets list "as a signal" when deciding which sites to exclude. They also cite a July 2026 post by Elon Musk saying he had asked the team to preserve rare books in a library and scan them "the hard way vs just cutting off the spine", read as confirmation that the company acquires and scans physical books.
No response had been filed at the time of reporting. These are allegations; no court has found the company liable.
Quellen
- Publishers sue Musk's AI firm over book scanning — aivio.news (secondary)