Last week Debian developers began considering a general resolution over AI large language model (LLM) usage within the Debian project. They are now up to weighing five different proposals for how to permit or block AI/LLM usage in the Debian space.

The discussion proposal continues on this Debian General Resolution (GR) with the five proposals now coming down to:

Proposal A: "This proposal aims to expressly forbid any contributions to Debian written with the use or assistance of large language models (LLMs) or other generative AI tools."

Proposal B: "Allow AI-Assisted Contributions"

Proposal C: "Reject LLMs (generative "AI") as far as practical"

Proposal D: "Accept AI contributions for Debian specific work"

Proposal E: "Responsible Use of Generative AI"

Proposal C for rejecting LLMs "as far as practical" comes down to also encouraging other free software and open-source communities to reject LLMs but noting not everyone will heed their calling. Proposal D is for allowing AI usage on the Debian websites, packaging, resources, and other Debian-specific assets. Lastly the details from Proposal E to note for responsible GenAI use is not endorsing or prohibiting GenAI tools for work within the Debian Project and allowing it to be used as long as it's disclosed and exercising appropriate care. It will be interesting to see what route the Debian developers ultimately decide to take around AI/LLMs.

Those curious about the details on each of these proposals being considered over generative AI / LLM usage within the Debian world can be found via the Debian Wiki.