Generative AI (LLM) Policy

Generative AI tools using large language models (LLMs) are widespread in software development today. While we acknowledge that these tools do have uses cases, LLMs engender harms to the pillars of Asahi Linux: software freedom, the open-source community, the open Internet, the environment, digital privacy, user consent, and software developers themselves. For more information on the ethical issues of LLM use in open source, we recommend reading SourceHut’s policy

Furthermore, Asahi Linux relies on clean room reverse engineering to ensure our work is legal. We have strict guardrails around binary disassembly and decompilation, and we absolutely forbid the use of leaked materials. While general questions around LLMs and copyright remain unsettled, LLMs pose unique legal risks to reverse engineering projects, as these systems are likely to violate the clean room requirements and taint the resulting code. This risk is amplified with “agentic” approaches, where the human may be unaware of the reverse engineering process. In reverse engineering and science how knowledge is obtained is just as important as the knowledge itself.

Due to these legal and ethical issues, we broadly forbid the use of generative AI tooling for material contributions to Asahi Linux. Enforcement may vary depending on the seriousness of the infraction. A GitHub issue drafted by an LLM may simply be closed with reference to this policy. A contributor using an LLM to interpret a trace made using the m1n1 hypervisor may be issued with a first and final warning. A contributor found to have concealed extensive LLM use will be banned immediately, particularly if they may have accessed unreleased Apple material.