why this existsMission
The association exists so that artificial intelligence belongs to everyone. The principles below come from its draft constitution, which explains its reasons rather than only stating rules.
the starting pointIntelligence is a means of production
Artificial intelligence is a tool, like a machine in a workshop. It can help people understand and learn, create, lighten their work, protect themselves and organise. A tool that matters this much should not belong to a handful of companies.
By workers, the project means every human being, whatever their job, status or ability.
the ratchetCopyleft keeps the commons open
Everything the association publishes is under a strong copyleft: today AGPL-3.0-or-later. Anyone who improves a model and serves it to the public has to publish their changes, so every improvement falls back into the common pot.
Open means, at least, the weights of the model, all the training code, and every source with its license, the date it was collected and checksums that let anyone verify it.
when in doubt, it is refusedPrinciple before performance
A text enters the training corpus only when its license is clear. When a license is doubtful, the text stays out. A clean corpus is worth more than a stronger model built on doubtful ground, because trust is lost faster than it is won.
No distillation: models are trained from scratch, with no output of another company's model in the data.
what is paid forThe service is sold, never the intelligence
Hosting a model costs money, and a paid service may one day cover that cost. What it sells is reliability, limits, support and infrastructure. The model itself stays free, self-hosting stays possible, and anyone may serve it too, for money if they like. That keeps prices honest.
Nobody gets rich from it: no profit goes to investors or to the people who run the association, and any surplus goes back to the mission.
a positionThe gains of automation belong to workers
Automation of work will happen. The association does not try to stop it. It argues that what automation gains should go to the people who work rather than to those who own the capital.