Post by @resucutie@plasmatrap.com

2026-09-21 23:30 UTC

@alexia@shrimp.starlightnet.work i dont know the relationship between amnesty international and KDE eco, but i assume the values align enough to also be taken in consideration from what i saw on their stance, i assume that an hypotetic AI model that has no awareness or "self consciousness" and its only ability being to generate code, and that was trained only on morally legal data, would be fine then?

Replies (2)

@alexia@shrimp.starlightnet.work 2026-09-21 23:32

@resucutie@plasmatrap.com For me personally, still no, but that is not the position this project takes per-se. IF such a model comes to exist, one MIGHT consider amending or changing the policy. I consider it unlikely considering the various human rights abuses that are baked into the creation of models currently, though.

Open ##4798961
@witchgirls@woof.tech 2026-09-21 23:36

@resucutie@plasmatrap.com It needs to be: 1) Trained on data obtained with consent only 2) Tagged with appropiately compensated labour 3) Enabled mainly through green non-planet-burning energy 4) Not inherently tied to extremely evil fascist companies 5) Completely taken out of mass surveillance projects 6) Only implemented while taking consent into consideration (i.e not forced into anyone) 7) Legally incapable of being used to completely replace humans in anything I don't really think that's ever going to happen though

Open ##4806819