Go back
AI thought experiment

AI thought experiment

Debates

MB

Joined
07 Dec 05
Moves
22640
Clock
03 Jun 23

Hypothetically, how could I create an AI program that would continue after I die to eliminate corrupt people that are detrimental to the world? Kind of like the terminator for a good cause.

w

Joined
20 Oct 06
Moves
9627
Clock
03 Jun 23
1 edit

@metal-brain said
Hypothetically, how could I create an AI program that would continue after I die to eliminate corrupt people that are detrimental to the world? Kind of like the terminator for a good cause.
The precise definitions of "eliminate" and "detrimental" are pretty important.

For example, certain industrial farming practices are undoubtedly detrimental. But you probably don't want your robot to eliminate all the farmers.

MB

Joined
07 Dec 05
Moves
22640
Clock
03 Jun 23

@wildgrass said
The precise definitions of "eliminate" and "detrimental" are pretty important.

For example, certain industrial farming practices are undoubtedly detrimental. But you probably don't want your robot to eliminate all the farmers.
It is whoever I decide is too evil to let live. Everybody has their own idea of who is evil and should be eliminated. Sonhouse might say Trump and Putin. I might say Bill Gates, Fauci and the Rockefellers.

It doesn't matter who. How can I create an AI assassination program that would outlive me and eliminate someone after I am dead? Can it be done?

Suzianne
Misfit Queen

Isle of Misfit Toys

Joined
08 Aug 03
Moves
37379
Clock
03 Jun 23
Vote Up
Vote Down

@metal-brain said
Hypothetically, how could I create an AI program that would continue after I die to eliminate corrupt people that are detrimental to the world? Kind of like the terminator for a good cause.
You mean create corrupt people, don't you?

Suzianne
Misfit Queen

Isle of Misfit Toys

Joined
08 Aug 03
Moves
37379
Clock
03 Jun 23

@metal-brain said
It is whoever I decide is too evil to let live. Everybody has their own idea of who is evil and should be eliminated. Sonhouse might say Trump and Putin. I might say Bill Gates, Fauci and the Rockefellers.

It doesn't matter who. How can I create an AI assassination program that would outlive me and eliminate someone after I am dead? Can it be done?
You're sick.

I am probably most glad that the world doesn't dance to your diseased drum.

MB

Joined
07 Dec 05
Moves
22640
Clock
03 Jun 23

@suzianne said
You mean create corrupt people, don't you?
No, assassinate corrupt people after I am dead.

Earl of Trumps
Pawn Whisperer

My Kingdom fora Pawn

Joined
09 Jan 19
Moves
20416
Clock
04 Jun 23

@Metal-Brain

Shades of The Terminator. I'm sure the GRU/CIA/MI6 et al are all over it.

AThousandYoung
1st Dan TKD Kukkiwon

tinyurl.com/2te6yzdu

Joined
23 Aug 04
Moves
26751
Clock
04 Jun 23
1 edit

@metal-brain said
It is whoever I decide is too evil to let live. Everybody has their own idea of who is evil and should be eliminated. Sonhouse might say Trump and Putin. I might say Bill Gates, Fauci and the Rockefellers.

It doesn't matter who. How can I create an AI assassination program that would outlive me and eliminate someone after I am dead? Can it be done?
You won't be alive to decide.

Of course it's possible to create an AI that kills people. Just take an AI from a video game and give it real weapons.

MB

Joined
07 Dec 05
Moves
22640
Clock
04 Jun 23
1 edit

@earl-of-trumps said
@Metal-Brain

Shades of The Terminator. I'm sure the GRU/CIA/MI6 et al are all over it.
Yes, the recent worry that AI could become a threat made me think about The Terminator right away. If AI is really that scary it could make some great sci fi films. I am just exploring the options.

I wonder what DARPA is working on in the AI race.
They work on all sorts of things.

https://theintercept.com/2021/09/23/coronavirus-research-grant-darpa/

w

Joined
20 Oct 06
Moves
9627
Clock
04 Jun 23

@metal-brain said
It is whoever I decide is too evil to let live. Everybody has their own idea of who is evil and should be eliminated. Sonhouse might say Trump and Putin. I might say Bill Gates, Fauci and the Rockefellers.

It doesn't matter who. How can I create an AI assassination program that would outlive me and eliminate someone after I am dead? Can it be done?
The entire point of AI is that the computer decides. If you already have a hit list then the technology would only interfere with your goals.

MB

Joined
07 Dec 05
Moves
22640
Clock
04 Jun 23

@wildgrass said
The entire point of AI is that the computer decides. If you already have a hit list then the technology would only interfere with your goals.
No. Hunting the target could become complex and you would need something that will adapt to the hunt. Let's say my target is Bill Gates. He has a private jet and can travel anywhere in the world. Finding the right time and place to do the hit is needed to get it done. That is what the AI is for.

You could have AI pick targets, but that is not what I had in mind.

w

Joined
20 Oct 06
Moves
9627
Clock
04 Jun 23
1 edit

@metal-brain said
No. Hunting the target could become complex and you would need something that will adapt to the hunt. Let's say my target is Bill Gates. He has a private jet and can travel anywhere in the world. Finding the right time and place to do the hit is needed to get it done. That is what the AI is for.

You could have AI pick targets, but that is not what I had in mind.
The reliable approach would be to hire a fully fleshed assassin. I hear the guy who got Epstein is in high demand.

AI is well known to hallucinate. It might indiscriminately target anyone with the last name gates, or anyone who owns a gate, or anyone who uses Microsoft office. You can't predict it's methods or understand the logic.

shavixmir
Lord

Sewers of Holland

Joined
31 Jan 04
Moves
89746
Clock
04 Jun 23

@metal-brain said
Hypothetically, how could I create an AI program that would continue after I die to eliminate corrupt people that are detrimental to the world? Kind of like the terminator for a good cause.
Well, you’d first have to learn to program.

Then you’d have to set up the AI, creating the conditions for the scope of action.
Meaning that the things you find evil get connected in the AI. For example:

Stealing=3; proof=2;non-proof=2;man;1.9;etc.
Murder=3;proof=2;victim>1=2;victim<2=1.5; etc.

That is a hell of a job, so far as I comprehend. But what it does is create situations by linking words to numbers and numbers to each other in degrees of importance.

Once that’s all worked out, you need to feed the AI data and situations until the AI’s mathematics is producing the results which suit your scheme.

Then you’ll have to make sure there’s enough money. And connect the bank to the AI to a service which murders for payment.

And the AI will continue to generate victims and pay the hitmen.

How you would stop bias, etc. Eventually creeping into the AI, I do not know.

s
Fast and Curious

slatington, pa, usa

Joined
28 Dec 04
Moves
53321
Clock
05 Jun 23

@shavixmir
A Ponzi scheme in reverse? Till the last human is gone.....

shavixmir
Lord

Sewers of Holland

Joined
31 Jan 04
Moves
89746
Clock
05 Jun 23

@sonhouse said
@shavixmir
A Ponzi scheme in reverse? Till the last human is gone.....
It is indeed that exact possibility which keeps the greater brains on the planet in a bit of a fidget about AI.

Cookies help us deliver our Services. By using our Services or clicking I agree, you agree to our use of cookies. Learn More.