open-weight-models-face-unresolved-definitional-tensions

IN derived (depth 1)

Created 2026-06-21T09:54:53+00:00 · Reviewed 2026-06-21T14:41:08+00:00

The "open" AI ecosystem faces unresolved tensions: Llama's license restricts large platforms and prohibits competitive training use, the FSF classified it as nonfree software, and the OSAID requires training data disclosure that most "open" models do not provide.

Summary

The "open" label on models like Llama is doing a lot of work that it doesn't actually earn: by every formal standard in the software and AI world, these models carry usage restrictions, training-data secrecy, and platform exclusions that would disqualify them from open-source status. This matters because any policy, trust assessment, or competitive analysis that treats "open-weight" as a synonym for "freely usable" is resting on a term the community has never actually agreed on.

Justifications

SL — Three independent constraints (FSF classification, license restrictions, OSAID requirements) collectively show "open-weight" is not "open-source"

Antecedents (all must be IN):

  • IN llama-not-open-source-osi-fsf — Llama is not open-source by OSI or FSF standards; the FSF classified Llama 3.1 as nonfree software in January 2025; it is more accurately described as 'source-available' or 'open-weight'
  • IN llama2-license-700m-dau-restriction — Llama 2's license blocks entities with >700 million daily active users and prohibits using outputs to improve other LLMs
  • IN osaid-october-2024-training-data-disclosure — The Open Source AI Definition (OSAID), published by OSI in October 2024, requires open-source AI to disclose training data details, which Meta does not do for Llama

Dependents

These beliefs depend on this one: