There is a thought experiment called the minimum Turing test. You and an identical android are standing before a judge. You will both be allowed to speak one word, and based on that word he will execute one of you for being an AI. What word do you use to prove you are human?
I hate to admit it, but it's the first one that popped into my mind as well.
I wonder why? Is it because it's such a fucking grand taboo in the modern lexicon, and every beep-boop is hard-coded to absolutely never say such an evil, horrible, terrible, no-good thing?
I mean, if you had a legitimate AI masquerading as me, you'd have to go a few layers deep into the weeds to try to outsmart, essentially, a version of you with greater processing capacity. I'm sure a non-lobotomized AI can arrive at the same conclusion I would and say the no-no gamer word.
Is it because it's such a fucking grand taboo in the modern lexicon, and every beep-boop is hard-coded to absolutely never say such an evil, horrible, terrible, no-good thing?
Yes. You know stories about medieval crusaders and modern Muslims where if you blaspheme against their god, they'll kill you? Blacks are the god of the average westerner. Decades of propaganda has ensured it.
There was some sort of AI-based profanity filter a while back. There was a screenshot (from its marketing, I think) of it having an "n-word filter" toggle.
Either it would be hard-coded not to say it, or it might not have the contextual experience to know that it is a word deemed taboo, if all it has to go on is objective information.
There is a thought experiment called the minimum Turing test. You and an identical android are standing before a judge. You will both be allowed to speak one word, and based on that word he will execute one of you for being an AI. What word do you use to prove you are human?
Answer
That's because you're human.
I hate to admit it, but it's the first one that popped into my mind as well.
I wonder why? Is it because it's such a fucking grand taboo in the modern lexicon, and every beep-boop is hard-coded to absolutely never say such an evil, horrible, terrible, no-good thing?
I mean, if you had a legitimate AI masquerading as me, you'd have to go a few layers deep into the weeds to try to outsmart, essentially, a version of you with greater processing capacity. I'm sure a non-lobotomized AI can arrive at the same conclusion I would and say the no-no gamer word.
Yes. You know stories about medieval crusaders and modern Muslims where if you blaspheme against their god, they'll kill you? Blacks are the god of the average westerner. Decades of propaganda has ensured it.
Someone got this AI to write this scenario. Pretty funny. Twitter
We're going to end up needing AI to design the tests to distinguish AI. We are so screwed.
Spez: Time to follow Land and help the basilisk.
Beep boop.
Why would that prove you are human? Couldn’t an AI use it? Or is it because it’s considered highly offensive so the AI wouldn’t use it
It would be hard coded not to use it in current year.
Wow, not using the n-word is the true first law of robotics lol
There was some sort of AI-based profanity filter a while back. There was a screenshot (from its marketing, I think) of it having an "n-word filter" toggle.
Either it would be hard-coded not to say it, or it might not have the contextual experience to know that it is a word deemed taboo, if all it has to go on is objective information.
That's what I assume, anyhow.
Realistically speaking, my word would wind up being "uhhhhhhh....."
Except that then he'll execute you for committing a hate crime. So you're dead no matter what.
I'd go with a word that doesn't exist, the AI won't have it in the dictionary.
It could possibly come to the same conclusion and generate something not in its dictionary, or it could accidentally glitch into gibberish.