Project Panama is a digital piracy and destructive book scanning operation by Anthropic aimed at developing a dataset of published old and rare books. Project P…
Project Panama is a digital piracy and destructive book scanning operation by Anthropic aimed at developing a dataset of published old and rare books. Project Panama is described in unsealed court filings from a 2024 class-action copyright lawsuit. In an internal planning document, Project Panama is described as Anthropic's "effort to destructively scan all the books in the world".[1][2][3]
Project Panama aims to produce a digital library for training through two methods.
Anthropic began by obtaining millions of previously digitized works through digital piracy. According to court documents, Anthropic used Books3, LibGen and Pirate Library Mirror for this purpose, and may have acquired up to 7 million books this way.[3][2]
In 2025, Anthropic settled with Libgen and Pirate Library mirror for US$1.5 billion for pirating 482,460 books and agreed to destroy any remaining copies of said books.[3] Authors usually received US$3,000 per book from this settlement.[3]
Anthropic purchased millions of used and rare books from online retailers such as Better World Books, and World of Books, and Zoom Books, sliced off their spines, and scanned their pages in order to train Claude.[1][4][2][5] This process involved the use of hydraulic cutters for spine removal and high speed scanner. The paper was then recycled.[5]
The dataset was then kept private and only used for training models.[2]
According to the Project Panama planning document, Anthropic did not "want it to be known that [it was] working on this".[1][2][5]
In the same 2025 settlement, Judge William Alsup ruled that the destruction and digitization of legally purchased books constituted fair use, in contrast to Anthropic's prior use of pirated copies.[1][4]
Informasi ini disarikan dari Wikipedia dan disajikan kembali untuk tujuan edukasi. Konten tersedia di bawah lisensi CC BY-SA 3.0. Kami tidak bertanggung jawab atas ketidakakuratan data yang bersumber dari kontribusi publik tersebut.