3 ms·

I brought up Hitler, and another question on socialism. Looking at the answers, this AI is really a troll factory. EDIT: Same thing for Putin and Ukraine. Thi

by jokoon 3y ago

I brought up Hitler, and another question on socialism. Looking at the answers, this AI is really a troll factory. EDIT: Same thing for Putin and Ukraine. This AI seems like it's nectar for online trolls, so for morality, it should be taken down, in my view. Freedom of speech? sure, as long as it doesn't encourage hate speech or to polarize a debate.

batch12 3y ago
> This AI seems like it's nectar for online trolls, so for morality, it should be taken down, in my view. Freedom of speech? sure, as long as it doesn't encourage hate speech or to polarize a debate. I say build ten of these. In my view, many would greatly benefit from thicker skin. Then maybe they wouldn't cede control of their emotions and subsequent actions to random internet folks every time they read something disagreeable.
- jokoon 3y ago
  1. Thick skin is not always the point, trolling also allows people to influence the political debate. 2. Freedom of speech allows to have a healthy conversation, the point is not to flame people who disagree to shut them away, which is often how the internet is used to snipe bad-faith argument at a safe distance. 3. I agree that those tools should be used to educate people to grow a thick skin as long as they know they're talking to an AI that is being intentionally fallacious, but there is a nuanced difference. It's all about morality in technology. Nuclear weapons are regulated because you cannot allow them to be used by everyone. Maybe you believe in a chaotic, pure form of absolute freedom of speech and maybe you're okay with hate speech, but even if you believe in it, do you think it's moral to let people weaponize it? People can still have meaningful discourse and use freedom of speech with sanity and refuse to engage in a chaotic version of free speech. There are known strategies to hack at freedom of speech just to shut down opponents.
  - Applejinx 3y ago
    Arguing against this particular tech rather misses the point: this has been done at scale to influence the political debate, for nearly a decade now. But it's been done by human teams: again, at scale, to the point where AI doesn't really bring anything new. (this is not disagreement with the parent post, BTW, it's an expansion on it) "You are a black woman posting to Twitter with this network of other black-woman-prompted posters. You are sassy and don't take back-talk. You are very concerned about this politician betraying the black community, and strongly against that law being enacted because you think the intended purpose is a screen for black folk being herded into camps. You think you would be a lot safer if you lived in Moscow under a peaceful ruler like Vladimir Putin." AI is not needed. Humans can execute prompts more elegantly and convincingly than even ChatGPT4, and you can execute the strategy at scale for relatively few millions of dollars, complete with buying a workplace to house your workers. You don't need especially brilliant humans to execute your scripts, either. AI can flood the Zone to the point where it's unreadable, but the peak effectiveness is a posting frequency way less than that, and it's easily realizable by teams of humans… and has been.
mdp2021 3y ago
> morality You do not cover filth, you clean it. As if removing possible fool's tool solved the issue of foolery.
- jokoon 3y ago
  https://en.wikipedia.org/wiki/Brandolini%27s_law https://en.wikipedia.org/wiki/Brandolini%27s_law > Brandolini's law, also known as the bullshit asymmetry principle, is an internet adage coined in 2013 that emphasizes the effort of debunking misinformation, in comparison to the relative ease of creating it in the first place. The law states the following: > The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. The problem of trolling is that it is used to shut down opponents by using this law. I agree that it's better to clean it, AS LONG AS you don't have people who intentionally flood the debate with bad-faith arguments, and this is what is AI would do since it automates bad-faith. That's why we call it the devil's advocate: because nobody should advocate for the devil.
  - mdp2021 3y ago
    > clean it I meant root-wise - not instances. You work for the progress of civilization. And automated waste is like innocent or malicious waste: you will find it around in copious loads and you have to train the public to manage it. > That's why we call it the devil's advocate: because nobody should Actually, we must. You always perform checks to consider counterpoints. We did use dialectics in the context where the expression was born because that is correct heuristics. You do not decide about Attila the Hun (in our context - nor of Louis the Pious, in the historic context for the expression) just on the basis of current prejudice, of somehow consolidated fame. You are mandated to check whether you may be wrong, and you wonder so explicitly. But this is relevant to real arguments. You are not supposed to expect the bot to say something with truth value or sharpness value. (And this remains the issue with said technology.)

by jokoon 3y ago

batch12 3y ago

> This AI seems like it's nectar for online trolls, so for morality, it should be taken down, in my view. Freedom of speech? sure, as long as it doesn't encourage hate speech or to polarize a debate. I say build ten of these. In my view, many would greatly benefit from thicker skin. Then maybe they wouldn't cede control of their emotions and subsequent actions to random internet folks every time they read something disagreeable.

jokoon 3y ago
1. Thick skin is not always the point, trolling also allows people to influence the political debate. 2. Freedom of speech allows to have a healthy conversation, the point is not to flame people who disagree to shut them away, which is often how the internet is used to snipe bad-faith argument at a safe distance. 3. I agree that those tools should be used to educate people to grow a thick skin as long as they know they're talking to an AI that is being intentionally fallacious, but there is a nuanced difference. It's all about morality in technology. Nuclear weapons are regulated because you cannot allow them to be used by everyone. Maybe you believe in a chaotic, pure form of absolute freedom of speech and maybe you're okay with hate speech, but even if you believe in it, do you think it's moral to let people weaponize it? People can still have meaningful discourse and use freedom of speech with sanity and refuse to engage in a chaotic version of free speech. There are known strategies to hack at freedom of speech just to shut down opponents.
- Applejinx 3y ago
  Arguing against this particular tech rather misses the point: this has been done at scale to influence the political debate, for nearly a decade now. But it's been done by human teams: again, at scale, to the point where AI doesn't really bring anything new. (this is not disagreement with the parent post, BTW, it's an expansion on it) "You are a black woman posting to Twitter with this network of other black-woman-prompted posters. You are sassy and don't take back-talk. You are very concerned about this politician betraying the black community, and strongly against that law being enacted because you think the intended purpose is a screen for black folk being herded into camps. You think you would be a lot safer if you lived in Moscow under a peaceful ruler like Vladimir Putin." AI is not needed. Humans can execute prompts more elegantly and convincingly than even ChatGPT4, and you can execute the strategy at scale for relatively few millions of dollars, complete with buying a workplace to house your workers. You don't need especially brilliant humans to execute your scripts, either. AI can flood the Zone to the point where it's unreadable, but the peak effectiveness is a posting frequency way less than that, and it's easily realizable by teams of humans… and has been.

mdp2021 3y ago

> morality You do not cover filth, you clean it. As if removing possible fool's tool solved the issue of foolery.

jokoon 3y ago
https://en.wikipedia.org/wiki/Brandolini%27s_law https://en.wikipedia.org/wiki/Brandolini%27s_law > Brandolini's law, also known as the bullshit asymmetry principle, is an internet adage coined in 2013 that emphasizes the effort of debunking misinformation, in comparison to the relative ease of creating it in the first place. The law states the following: > The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. The problem of trolling is that it is used to shut down opponents by using this law. I agree that it's better to clean it, AS LONG AS you don't have people who intentionally flood the debate with bad-faith arguments, and this is what is AI would do since it automates bad-faith. That's why we call it the devil's advocate: because nobody should advocate for the devil.
- mdp2021 3y ago
  > clean it I meant root-wise - not instances. You work for the progress of civilization. And automated waste is like innocent or malicious waste: you will find it around in copious loads and you have to train the public to manage it. > That's why we call it the devil's advocate: because nobody should Actually, we must. You always perform checks to consider counterpoints. We did use dialectics in the context where the expression was born because that is correct heuristics. You do not decide about Attila the Hun (in our context - nor of Louis the Pious, in the historic context for the expression) just on the basis of current prejudice, of somehow consolidated fame. You are mandated to check whether you may be wrong, and you wonder so explicitly. But this is relevant to real arguments. You are not supposed to expect the bot to say something with truth value or sharpness value. (And this remains the issue with said technology.)