How Dangerous is AI?

[blueone]

An interesting commentary by Eric Schmidt, Henry Kissinger, and Daniel Huttenlocher in the Wall Street Journal. Apologies that it probably requires a subscription, but if you can it's worth reading.

Opinion | ChatGPT Heralds an Intellectual Revolution

Generative artificial intelligence presents a philosophical and practical challenge on a scale not experienced since the start of the Enlightenment.

www.wsj.com

 

[Hugo9000]

We were warned by 2001. Luckily, Star Trek taught us how to disable AIs by asking them to calculate pi to all decimal places.

There's also the method used in Fredric Brown's 1942 classic short story, ETAOIN SHRDLU. It was about a Linotype machine that achieved sentience. I won't include a spoiler, since it's one of my favorite short stories haha! As far as I know, it's the first cautionary tale about artificial intelligence.

The story's title comes from the keys on the Linotype, arranged in descending order of frequency in the English language (as determined by experienced manual compositors of the time, which differs from other letter frequency rankings such as those used in cryptography).

 

[jsilvela]

Long before there is a real danger of AI's wiping out humanity there will be massive disruption of the job market. I saw one prediction that lawyers were particularly vulnerable, but I believe the author did not really understand what lawyers do. However, I believe paralegals are at risk.

That's a great point.

Even if the AI we have right now is not really "intelligence", it can and will disrupt the job market.
This has been the case with computers and software for decades already.
And before then with all forms of mechanized labor.
The conventional economic/political wisdom has been that on aggregate, there was not net job destruction.

I've seen some economists start to revise the doctrine. Some are talking of net job destruction.
And, predictably, the fearful politicians and journalists and CEO's are cropping up only now that they start fearing for the destruction of white collar jobs.

 

[MRC01]

That's a great point.
Even if the AI we have right now is not really "intelligence", it can and will disrupt the job market.

I agree. Historically, when technology is applied to the real world—from fire to steam engines to aviation to digital communication—it always creates better quality products and services and more productive, higher paying jobs. Everyone benefits in the long term. Yet in the short term it can cause significant economic dislocation as people and the economy adjust to the changes. AI has the potential to cause the biggest disruption in our generation, due to its rapid growth and applicability across widespread areas of activity.

 

[fpitas]

Still, unless they come up with ways to make AIs more "falsehood resistant", using one to replace a person will always be a dodgy affair.

 

[FrantzM]

Our societies are predicated upon Humans being at the top. Machines work for us. We design these for our purposes. This new technology has the potential to change this paradigm. Who will work for whom? then?
We keep on focusing on the mistakes those AIs make, so far these may have been innocuous. Based on our current complete dependency on machines, yes , people, The potential of such falsehoods being harmful, is not to be taken lightly. We focus on their current weaknesses, yeah , they sprout nonsenses but they learn, continuously with no rest or seconds off, they shall become better.

Peace.

 

[kemmler3D]

Long before there is a real danger of AI's wiping out humanity there will be massive disruption of the job market.

Strong agree on this.

Is it possible to give an AI a moral compass or true empathy?

Moral compass? Yes. Morals are broad rules about what's right or wrong. TRUE Empathy? Probably not, this implies an emotional connection, and we know how to give machines rules, but not feelings. Simulated empathy? Probably, machines would just need to be able to simulate human mental states in some way, and then have morals applied to those simulations.

I think getting these things wrong would have dire consequences, though.

 

[kemmler3D]

The idea I had in mind was the butterfly effect. If you can influence a quantum probability, that puts an electron in THIS state rather than THAT state, this affects the next atom, and so on. It cascades into macro effects.

I suppose, but what exactly is exerting this influence that we might call "will"?

 

[antcollinet]

This, powered by GPT-4, what can go wrong.

View attachment 276195

You have 20 seconds to comply...

 

[Timcognito]

2020s ChatGPT

 

[MRC01]

I suppose, but what exactly is exerting this influence that we might call "will"?

Great question. I'll paraphrase Justice Potter Stewart: I can't define it, but I know it when I experience it. All humans perceive that we have it. That doesn't necessarily mean it is what we perceive it to be, but the experience is compelling enough and survives test & observation, so its existence cannot be dismissed.

 

[fpitas]

View attachment 276236
2020s ChatGPT

The internet kicked in right around 2000, and IQs started plummeting.

 

[Count Arthur]

You have 20 seconds to comply...

 

[egellings]

When the AI can empty your bank account and not be traceable, I'd call it dangerous.

 

[fpitas]

When the AI can empty your bank account and not be traceable, I'd call it dangerous.

Can Scambot AI be far behind?

 

[computer-audiophile]

The internet kicked in right around 2000, and IQs started plummeting.

It was some years earlier in 1993 that the World Wide Web really became popular and widely used with the introduction of the Netscape browser.

 

[fpitas]

It was some years earlier in 1993 that the World Wide Web really became popular and widely used with the introduction of the Netscape browser.

Yes, but then and for some years it was only us boffins. The internet was used for research, and to perfect the tracing of female anatomy using ASCII characters.

 

[Ron Texas]

How much of this doom from AI prediction is rooted in the recent interest or belief in a coming apocalypse? After all there are endless movies and TV series about zombies, the latest of which is "The Last of Us".

What if an AI trained on the Georgia Guidestones inscription: Maintain humanity under 500,000,000 in perpetual balance with nature.

 

[fpitas]

How much of this doom from AI prediction is rooted in the recent interest or belief in a coming apocalypse? After all there are endless movies and TV series about zombies, the latest of which is "The Last of Us".

What if an AI trained on the Georgia Guidestones inscription: Maintain humanity under 500,000,000 in perpetual balance with nature.

When the conspiracy blows up big, we know it started here

 

[antcollinet]

What if an AI trained on the Georgia Guidestones inscription: Maintain humanity under 500,000,000 in perpetual balance with nature.

Utopia

Apart from the bit about getting from around 8 down to 0.5 billion . Of course the biggest habitual consumers would have to go first.