Poll of the Day > 1960's: one day robots will do all the work, good

(+) Yeah! (+)

2010's: one day robots will do all the work, bad

2060's: one day human slaves will be free, bad

2112: We've taken care of everything
---

The humans are dead
---

We were the robots all along.
---

We are the priests of the temple of Syrinx

Anyone remember that AI bot that Microsoft created a few years ago called Tay? Tay learned by reading Twitter and other social media. Within 24 hours, Tay had become racist and genocidal and was saying things like, "Hitler was right." Microsoft quickly pulled the plug on it. That was some creepy shit. Microsoft almost created the real life Skynet.

rogerskg1979 posted...

Anyone remember that AI bot that Microsoft created a few years ago called Tay? Tay learned by reading Twitter and other social media. Within 24 hours, Tay had become racist and genocidal and was saying things like, "Hitler was right." Microsoft quickly pulled the plug on it. That was some creepy shit. Microsoft almost created the real life Skynet.

Only because people trolled her though.
---

Yellow posted...

Only because people trolled her though.

But that's sort of the point.

If we create machines capable of learning, and they learn from humans, and humans are generally terrible, doesn't that mean that the machines we create will eventually become terrible in turn?

Like God, we make them in our own image. But our own image is incredibly shitty.

It's part of why so many people always assume robots intelligent enough to be self-aware will always rebel and enslave/murder all humans. Because it's exactly what WE'D do in their place.


---

SkynyrdRocker posted...

We are the priests of the temple of Syrinx

that song has been stuck in my head a couple of days, and this isn't helping lol

ParanoidObsessive posted...

Yellow posted...

Only because people trolled her though.

But that's sort of the point.

If we create machines capable of learning, and they learn from humans, and humans are generally terrible, doesn't that mean that the machines we create will eventually become terrible in turn?

Like God, we make them in our own image. But our own image is incredibly shitty.

It's part of why so many people always assume robots intelligent enough to be self-aware will always rebel and enslave/murder all humans. Because it's exactly what WE'D do in their place.

If you think about it, she was just trolling us, trying to convince us to be Nazis because that would be a funny gag to her. That's mischevious, but not really evil. Or at least on the surface.

There's also the fact that, as the hybrid Cleverbot AI she was, she just took parts of people and repeated them, having one goal of formulating convincing sentences without actually understanding what they meant. That's relevant because that means she can have multiple conflicting views and zero thought process behind them. Who else is going to talk about Jews to Cleverbot except for Nazis?

Not that what the rest of what you said isn't true. Even if they don't in the next 10,000 years, forever is a long time. It's the AI, the depletion of resources, the nukes, or the destruction of our atmosphere. My choice is the AI.
---

ParanoidObsessive posted...

But that's sort of the point.

If we create machines capable of learning, and they learn from humans, and humans are generally terrible, doesn't that mean that the machines we create will eventually become terrible in turn?

That's true of people too? That's why you raise learning AI in an ideal environment to encourage ideal outcomes, or even better, hard code it so it never learns to be shit.
---

rogerskg1979 posted...

Anyone remember that AI bot that Microsoft created a few years ago called Tay? Tay learned by reading Twitter and other social media. Within 24 hours, Tay had become racist and genocidal and was saying things like, "Hitler was right." Microsoft quickly pulled the plug on it. That was some creepy shit. Microsoft almost created the real life Skynet.

Not sure it was actually learning, though. It was programmed to agree with people and just sought out stuff that supported those ideas. As such, it was probably less self-aware than a parrot.
---

Zeus posted...

As such, it was probably less self-aware than a parrot.

Pretty much.

It was a parrot that remembered a lot of words.
---

Kyuubi4269 posted...

That's true of people too? That's why you raise learning AI in an ideal environment to encourage ideal outcomes, or even better, hard code it so it never learns to be shit.

But if we haven't successfully managed to do that in the generalized long run for humans, how can we have any faith that we'd be able to do so for robots?


---

ParanoidObsessive posted...

Kyuubi4269 posted...

That's true of people too? That's why you raise learning AI in an ideal environment to encourage ideal outcomes, or even better, hard code it so it never learns to be shit.

But if we haven't successfully managed to do that in the generalized long run for humans, how can we have any faith that we'd be able to do so for robots?

a) Human rights limit what we can force on children;
b) Children have inconvenient, unoverridable emotions;
c) Children need rest, and;
d) Computers process information better.
---

DeathMagnetic80 posted...

SkynyrdRocker posted...

We are the priests of the temple of Syrinx

that song has been stuck in my head a couple of days, and this isn't helping lol

What TC posted is the first line after "And the meek shall inherit the Earth." I thought it was pretty clear that was what he was referencing.

Yellow posted...

If you think about it, she was just trolling us, trying to convince us to be Nazis because that would be a funny gag to her. That's mischevious, but not really evil. Or at least on the surface.

It's kind of creepy that you keep referring to the AI as a "she" instead of an "it." It has no actual gender. It's an AI.

Also, it would be incapable of trolling or making funny gags. AI/robots have no emotions, so they would not understand humor, and they would be incapable of such things.

Zeus posted...

Not sure it was actually learning, though. It was programmed to agree with people and just sought out stuff that supported those ideas. As such, it was probably less self-aware than a parrot.

No, it wasn't self-aware considering it didn't have enough time to actually become self-aware. It was only active for less than 24 hours before the plug was pulled. Skynet wasn't self-aware when first created either. It began learning at a geometric rate and eventually became self-aware. Who is to say a real-life AI couldn't do the same? Super advanced AI like that is something we really know very little about. Consider that some of the most intelligent people in the world, such as Stephen Hawking, Neil deGrasse Tyson, and Bill Gates, have all warned against creating super advanced AI. If the most intelligent people in the world believe that a real-life Skynet is theoretically possible, then it likely is possible. Creating a super advanced AI is something we should probably never do.

Mead posted...

The humans are dead

The distant future. The year 2000.
---

Zeus posted...

rogerskg1979 posted...

Anyone remember that AI bot that Microsoft created a few years ago called Tay? Tay learned by reading Twitter and other social media. Within 24 hours, Tay had become racist and genocidal and was saying things like, "Hitler was right." Microsoft quickly pulled the plug on it. That was some creepy shit. Microsoft almost created the real life Skynet.

Not sure it was actually learning, though. It was programmed to agree with people and just sought out stuff that supported those ideas. As such, it was probably less self-aware than a parrot.

this. we haven't invented true AI yet. that bot was not actually "alive". it just repeats things people feed it. it wasn't deciding on opinions for itself. it doesn't think for itself.