I got some feedback privately to clarify some more things.
---updated proposal v4.0 ---Using its power under Constitution section 4.1 (5), the project issues the following statement describing its current position on AI-assisted contributions. This statement describes the position of the project at the time it is adopted. That position may evolve as time passes without the need to resort to future general resolutions. The GR process remains available if the project needs a decision and cannot come to a consensus.
PreambleLLMs and its impact on society and environment is a hugely polarizing debate.
In this document, we mainly look at contributions using LLMs.LLMs have made a huge leap in language translations, but at the same time depending on big tech companies to provide this service is not a good idea. When it comes to generating code, we also need to consider how it weakens the idea of copyleft, unlike translations.
Enforcing copyleft requires us to depend on copyright law, if LLM generated code cannot be copyrighted, then LLM could be used to take out copyleft requirements from Free Software projects. But this is not something we can really control.
Basics1. If a Free Software LLM model is run locally, we have some control over it (we still need to retrain it, if we want to modify it). But even for that level of control, the high hardware specifications required to run these models makes access to such hardware difficult. So we cannot entirely insist on running an LLM model locally, similar to not insisting on use of Free Hardware Designs. So we will need to collectively run the models to make sure everyone have access to Free Software models in practice.
2. Public AI is AI as public infrastructure like highways, water, or electricity. Public AI uses Free Software and Open Weight models. Open Weights refer to the final weights and biases of a trained neural network. These values, once locked in, determine how the model interprets input data and generates outputs.
Since Public AI is not a well defined term, we need a more precise definition. PublicAI.Co also complies with Open Source AI definition by OSI (https://opensource.org/ai) so we will base Public AI definition on Open Source AI definition.
For AI to be accessible to people, it has to be available easily to people, like Wikipedia or Lets Encrypt. Not for profit collaborations between multiple organizations and public funding is essential to keep it sustainable, though public funding is not a requirement. How we are going to fund such Public AI will evolve in the future.
So to evaluate an AI web service, we will use the definition of Open Source AI by OSI, but some requirements may be relaxed.
3. At this point, we don't want to insist on training data being freely available, similar to how we don't insist on Free Hardware Designs. The challenges to practically being able to use the available data should be factored in. Similar to how costly it would be to produce hardware from Free Hardware Designs, even when training data is available, huge computing power is required to use it practically.
This point is a specific relaxation from the Open Source AI definition. So Public AI = Open Source AI + relaxed requirement on training data. Guidelines 1. Translations generated by Public AI can be accepted.2. Code / documentation generated by Publc AI can also be accepted. Ideally it should preserve share alike / copyleft conditions of original code used in training, but how copyright law evolves is not in our control.
3. Contributions (code/bug reports/patches/documentation etc) assisted by Public AI should be accepted.
4. Any such contributions should also follow the generally accepted norms for contributions like ownership and quality.
The person submitting any such contributions generated or assisted by Public AI should take full responsibility for its quality.
5. Any contributions using generative AI/LLM systems not matching requirements of Public AI will not be accepted.
ReferencesAn example to understand what Public AI looks like in practice is https://publicai.co, which is funded by Mozilla, the Future of Life Institute, and the Center for Cultural Innovation. This project is the inspiration and basis for this document.
---updated proposal v 3.0 ---If you feel Open Source AI is missing in the current set of ballot options, I'd appreciate seconds for this option. We could still clarify the draft if there are ambiguities.
Summary: The core part of this proposal is, Accept if Public AI, else reject. Public AI = Open Source AI + relaxed requirement on training data.
OpenPGP_0x8F53E0193B294B75.asc
Description: OpenPGP public key
OpenPGP_signature.asc
Description: OpenPGP digital signature
