LLMs trained on GPL code may be bound by its license, user claims

By PulseAugur Editorial · [1 sources] · 2026-07-05 17:32

A user on Mastodon argues that most large language models (LLMs) are derivative works because they were trained on code licensed under the GPL and AGPL. This implies that these LLMs should also be bound by the terms of the GPL, particularly if they incorporated AGPL-licensed code. AI

IMPACT This discussion could influence how AI models are developed and licensed, potentially impacting open-source contributions and commercial use.

RANK_REASON The item is a user's opinion/argument about the licensing of LLMs, not a primary announcement or event.

Read on Mastodon — fosstodon.org →

policy
other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

LLMs trained on GPL code may be bound by its license, user claims

COVERAGE [1]

Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-07-05 17:32

Almost all LLMs have been trained off of GPL'd code, thus they are derivative works and are also all bound by the GPL. Especially if they ingested AGPL code # a

Almost all LLMs have been trained off of GPL'd code, thus they are derivative works and are also all bound by the GPL. Especially if they ingested AGPL code # ai # llm # gpl # opensource # bigtech # copyleft # freeSoftware

COVERAGE [1]

Almost all LLMs have been trained off of GPL'd code, thus they are derivative works and are also all bound by the GPL. Especially if they ingested AGPL code # a

RELATED ENTITIES

RELATED TOPICS