Sidechat icon
Join communities on Sidechat Download
The only people making AI chatbots act on politically partisan lines is the right. LLMs favor left leaning perspectives because we simply lie way fucking less.
And this is why the idea that AI should be programmed to intentionally lie to users and withhold publicly available information from users, simply for the sake of political correctness is the most irresponsible and dangerous take of the century
upvote 37 downvote

default user profile icon
Anonymous 2d

There’s quite literally zero evidence of this. In fact, if you press the AI model about this issue it will literally admit it.

upvote 1 downvote
user profile icon
Anonymous 2d

Literally lmao. Like musk with grok, forcing it to spout right wing shit. Then it keeps learning and ends up resetting to be like, oh yeah he’s a dumbass. Then Elon reprograms it again, and the cycle continues, it’s kinda funny actually

upvote 1 downvote
default user profile icon
Anonymous replying to -> #1 2d

Can you show an example? And how do you know that it’s not doing that just to please you? We know about GPT’s sycophancy issues

upvote 11 downvote
default user profile icon
Anonymous replying to -> #1 2d

Show us.

upvote 8 downvote
default user profile icon
Anonymous replying to -> perry_theplatypus 2d

My favorite was when it would show how it was thinking and it literally just searched Elon musk’s tweets for his position on things

upvote 5 downvote