It is impossible to create any system that cannot be corrupted. Especially if said system runs on opaque technology created in secret by a private entity using a nondeterministic process. Especially if the system itself was designed to be nondeterministic.
If you want to interject AI into the governmental process it’d be way better to separate governance of the legislative process away from legislators, since there’s all kind of chicanery that we don’t easily see. ( the majority party in either chamber picks what votes they want to have, decides when voting is done, and even decides what day it is and how each member is recorded as having actually voted. )
It is impossible to create any system that cannot be corrupted. Especially if said system runs on opaque technology created in secret by a private entity using a nondeterministic process. Especially if the system itself was designed to be nondeterministic.
If you want to interject AI into the governmental process it’d be way better to separate governance of the legislative process away from legislators, since there’s all kind of chicanery that we don’t easily see. ( the majority party in either chamber picks what votes they want to have, decides when voting is done, and even decides what day it is and how each member is recorded as having actually voted. )
What part of open source model didn’t you understand? And one created before the plan was announced?
“Open Source” LLMs are, by their very nature, about as open as a compiled EXE offered without source.