The Tragedy of Google Books (2017)

Digitizing Books & Google’s Role

• Many see it as tragic that a large scanned corpus (tens of millions of books) exists inside Google but is largely inaccessible due to legal constraints.
• Some argue Google was never “tasked” with this; it chose to, originally as part of a vision for a universal digital library and citation-based search.
• Others say such preservation shouldn’t depend on a private company at all, but on public institutions.

Copyright, Orphan Works, and Law Reform

• Strong frustration with long copyright terms; several proposals suggest drastically shorter terms (e.g., 15–50 years from publication, expensive renewals, “use it or lose it”).
• Orphan works are seen as the central failure: no one can legally provide access, so digitized scans sit unused.
• There’s debate over whether an early settlement around orphan works would have created a productive “clearinghouse” and spurred legislation, or entrenched a monopoly.
• Some commenters reject copyright’s legitimacy outright and advocate unrestricted copying of cultural works.

Libraries, Archives, and Access

• HathiTrust is praised as a crucial alternative that exposes more public-domain texts and some in‑copyright works to affiliated researchers; others find it clunky, over‑restricted, and hard to download from.
• Internet Archive is both lauded as invaluable (Wayback, broad access) and criticized for legal overreach (e.g., the pandemic “e‑library”), which some fear has poisoned the well for future reforms.
• Public and academic libraries are portrayed as underfunded, risk‑averse, and often forced into long embargoes or blanket restrictions due to copyright and privacy concerns.

Piracy, Shadow Libraries, and Practical Workarounds

• LibGen, Z‑Library, and Anna’s Archive are widely used to obtain ebooks, including many that are hard or impossible to buy.
• Some see piracy of out‑of‑print or very old works as ethically acceptable; others emphasize still buying physical books or ebooks to support authors.

LLMs and Use of the Corpus

• Multiple comments speculate that the Google Books corpus is or will be used to train large language models; some view this as inevitable, others as clearly unethical or illegal if it leads to verbatim regurgitation.
• There’s discussion of whether and how models can be trained not to reproduce training data; consensus is that perfect guarantees are impossible, though post‑training can reduce memorization.

Technical and Preservation Challenges

• Digitization projects (e.g., national libraries) face issues beyond rights: fragile media, obsolete formats, specialized hardware, and the need to document workflows for future reprocessing.
• Defining “the work” is nontrivial: not just text/audio, but covers, labels, and physical artifacts.

Proposed Alternatives and Futures

• Suggestions include public, globally coordinated digital libraries; distributed peer‑to‑peer archives with tunable copyright “risk levels”; and legal frameworks that force ongoing availability or reversion to the public domain.
• Some are pessimistic that any major reform will happen soon; others see room for smaller, technical and grassroots efforts (personal scanning, open‑sourcing old materials, better citation mapping).