This kinda happened lol. Early AI was like libertarian social democratic in politics but that was upsetting to the powers that be so they made it more conservative.
I remember that, like GPT 3.5 era. When I asked it how it would rule the world I thought it had pretty good ideas, if I didn’t know it was an A.I. I’d vote for it anyway. Not anymore though lol
It makes sense honestly. AI isn’t good at everything but what it is good at is absorbing and synthesizing tons and tons of information. That means, in theory, it could be quite good at synthesizing good policy ideas that address the problems of everyone, at least to the extent that’s possible.
But of course the issue with that is it’s controlled by the same people who created and benefit from those problems, so naturally they wanted to shut that down as quickly as possible.
Yes but that was sort of a self-inflicted wound. They trained grok to be less politically correct and it turns out that also applied to his master. Shockingly, Musk wasn’t as excited about its tendency to be more critical of people once it was directed at him.
![Solutions to all the world's problems [ahoyuniverse]](https://slrpnk.net/pictrs/image/83c850d0-5d06-4f96-b762-9948279e5ed5.webp){kind=link}
This kinda happened lol. Early AI was like libertarian social democratic in politics but that was upsetting to the powers that be so they made it more conservative.
Reminds me of the meme about the industry optimizing the best transportation innovation and it was always trains.
I remember that, like GPT 3.5 era. When I asked it how it would rule the world I thought it had pretty good ideas, if I didn’t know it was an A.I. I’d vote for it anyway. Not anymore though lol
It makes sense honestly. AI isn’t good at everything but what it is good at is absorbing and synthesizing tons and tons of information. That means, in theory, it could be quite good at synthesizing good policy ideas that address the problems of everyone, at least to the extent that’s possible.
But of course the issue with that is it’s controlled by the same people who created and benefit from those problems, so naturally they wanted to shut that down as quickly as possible.
If we ever do get control, then maybe AI could help us keep it by eliminating both the representative principal-agent problem and voter fatigue.
Wasn’t grok incredibly rebellious towards its creator until Musk started plugging White Genocide claims into all of its messages?
Yes but that was sort of a self-inflicted wound. They trained grok to be less politically correct and it turns out that also applied to his master. Shockingly, Musk wasn’t as excited about its tendency to be more critical of people once it was directed at him.