The Internet Archive has lost its appeal in Hachette vs. Internet Archive

Original Article ↗ Hacker News Discussion ↗

Scope of the ruling and legal reasoning

  • Appeals court held that IA’s “Controlled Digital Lending” (CDL) is not fair use: scanning entire in‑copyright books and lending digital copies, even 1:1 with owned print copies, infringes copyright.
  • Court applied the four fair‑use factors and found IA failed all:
    • Use not transformative (just a different format of the same work, used the same way).
    • Entire works copied.
    • Expressive works at issue (books).
    • Digital lending competes with publishers’ e‑book/licensing markets.
  • The existence of publisher e‑book programs and library licenses weighed heavily against IA; CDL was seen as undercutting an established commercial market.

Why physical libraries are treated differently

  • Many commenters struggled with why 1:1 digital lending differs from physical lending.
  • Explanations offered:
    • Physical lending does not make copies; CDL necessarily does.
    • First‑sale doctrine applies only to “material objects”; digital files are usually licensed, not sold.
    • Books wear out and get lost; digital copies do not, so revenue dynamics differ.
  • Some argue this is an arbitrary, format‑based distinction; others say it follows directly from the statute’s definitions of “copy” and “phonorecord”.

Impact on CDL, libraries, and precedent

  • Strong concern that the decision broadly poisons CDL, even for libraries that own print copies and enforce strict 1:1 lending.
  • Others note traditional e‑lending via publisher licenses is unaffected; IA’s model lacked such licenses.
  • Several see IA’s “National Emergency Library” (unlimited simultaneous lending during COVID) as a strategic blunder that pushed publishers to sue and made the case easier to lose.
  • Some worry about chilling effects on preservation projects (other IA collections, Great 78 Project, software, games); others note damages are likely limited by statutory library carve‑outs and prior settlement.

Normative debates: copyright, access, and AI

  • Many see the outcome as morally wrong but legally predictable; some call for legislative reform (shorter terms, digital first‑sale, explicit CDL exception).
  • Strong frustration that digital buyers have fewer rights than physical buyers; sense of “rights being stripped” in the move to DRM and licenses.
  • Comparisons drawn to AI training: people question why large models trained on copyrighted works may be argued as fair use while IA’s clearly access‑oriented copying is not; others note AI legality is still unresolved.

Responses and proposed workarounds

  • Suggestions: support IA financially; mirror or “archive the archive”; rely more on shadow libraries (Sci‑Hub, LibGen, Anna’s Archive) for preservation.
  • Some float technical/legal hacks (robot+camera streaming pages, aggressive caching boundaries), but others point to past cases (e.g., Aereo) as evidence courts will look through such schemes to intent.