LLM-gen-AI - Software Freedom Conservancy
Get the latest update on our Vizio court case -->
Close navigation menu
Until July 17ᵗʰ,
we'll direct1 newly initiated Sustainerships & SFC donations to our software<br>right-to-repair work & efforts to resolve Bambu Lab's AGPLv3 violations!
$163,534 raised!
$86,473 to go!
Help us reach this goal<br>so we<br>can dedicate long-term full-time SFC<br>staff on our software right to repair work for 3D printers, & take action on Bambu Lab's AGPLv3 violations!
Recommendations
When Using LLM-backed Generative AI<br>Systems for FOSS Contributions
Preamble
The entire community of computer users, which quickly approaches<br>every human, faces the growing conundrum of generative artificial<br>intelligence systems backed by Large Language Models (“LLM-gen-AI”)1. Software freedom activists face<br>particularly difficult challenges in this regard; these LLM-gen-AI<br>systems have been applied in earnest to the endeavors of software<br>creation and modification.
We cannot sufficiently mitigate this tricky problem with merely one<br>statement or a few blog posts. In 2022, Software Freedom Conservancy<br>began our journey on this particular issue when our policy fellow,<br>Bradley M. Kühn, published If<br>Software is My Copilot, Who Programmed My Software?. In the last<br>year, that journey grew in complexity and urgency when some of SFC’s<br>member projects and supporters began to regularly request moral and<br>ethical guidance on these matters. SFC spent these months in<br>almost-daily internal discussions about the plethora of dilemmas<br>presented by LLM-gen-AI systems.
In 2024, SFC<br>published an aspirational statement, a thought experiment rather<br>than a definition. We now make urgent recommendations to those ordered by<br>their employers to use LLM-gen-AI code assistants to contribute to Free<br>and Open Source Software (“FOSS”).
Some FOSS project leaders have taken a zero-tolerance approach to any<br>LLM-gen-AI contributions to their projects. We support leaders who make<br>such decisions. FOSS project leaders deserve our sympathy and<br>understanding regarding the volumous onslaught of new contributions.<br>Patch evaluation has always required careful analysis (after all, humans<br>write bad code too). Now, that analysis demand (reasonably) feels<br>daunting to maintainers. Everyone should respect their decisions.
Nevertheless, we cannot and must not ignore the many FOSS<br>contributors who decide to explore these tools for the betterment of<br>FOSS. Software freedom activism only succeeds when we admit that we are<br>at least decades away from universal software freedom. Proprietary<br>systems will continue to exist; there is a real danger they will<br>continue to leapfrog FOSS. We should resist the use of proprietary<br>systems, which include the most popular LLM-gen-AI systems, but we<br>should also remain willing (as<br>we always have) to utilize such systems when they can advance<br>software freedom.
After much study, consideration, collaboration, and consultation with<br>many FOSS leaders, SFC formulated the following recommendations for FOSS<br>contributors who have decided to use LLM-gen-AI systems to augment their<br>FOSS work. We expect to update these recommendations periodically. These<br>are not mandates, demands, conclusions, nor definitions; rather, they<br>are best practices that we have formulated after careful study of the<br>undeniable reality that some FOSS contributors do want to use these<br>LLM-gen-AI systems.
In the months following the announcement of these recommendations,<br>SFC plans an ongoing engagement campaign, including documents, online<br>tutorials, public Q&As, and other community engagement, on these<br>matters. SFC does not make these recommendations in isolation; rather,<br>we offer sustained assistance to the community, particularly to FOSS<br>projects working with proprietary LLM-gen-AI systems.
The long term goal of software freedom is to eliminate the harm of<br>proprietary technology. While we work toward that greater goal, we<br>should seek to mitigate the harms that we cannot immediately eliminate.<br>These recommendations aim to abate the damage of these systems, and also<br>consider how these tools might counter-intuitively help us<br>advance FOSS.
Recommendations
These recommendations are listed in order of our view of their<br>relative importance (most important first).
The FOSS community should support, not just tolerate,<br>those who outright reject LLM-gen-AI systems. There are many<br>intersecting ethical and moral issues regarding these systems, many of<br>which are not currently fully understood. Anyone who chooses to avoid<br>them deserves our support and assistance.
Every FOSS contributor deserves self-determination<br>regarding LLM-gen-AI. No one should be required to use<br>these systems under duress. We make special note here of the increasing<br>reports from technology workers who have been ordered by their<br>management (often under penalty of termination) to use these systems for<br>all their work: FOSS and proprietary. Such mandates are unconscionable<br>and we call on the...