FSF threatens Anthropic over infringed copyright: share your LLMs freely

8 hours ago

Anthropic was sued in 2024 for allegedly infringing copyrights while training its LLMs.
The Free Software Foundation (FSF) claims Anthropic's training data included their copyrighted book, 'Free as in Freedom: Richard Stallman's Crusade for Free Software', published under the GNU Free Documentation License (GNU FDL).
FSF urges Anthropic and other LLM developers to share complete training inputs, models, configurations, and source code with users to uphold computing freedom.
FSF, though resource-limited, may seek user freedom as compensation if their copyright is found violated in lawsuits like Bartz v. Anthropic.
The FSF's stance is summarized as not usually suing for copyright infringement but settling for freedom when they do.
Debate arises over the meaning of providing LLMs 'in freedom', with interpretations ranging from free usage by everyone to releasing models under free licenses.
Critics question the coherence of FSF's demands, while supporters highlight RMS's historically accurate 'extreme' predictions regarding software freedom.
The discussion touches on the implications of copyright laws and fair use in the context of AI model training and the potential for derived works to be considered under copyleft.

Hasty Briefsbeta