"Powerful" AI writes essay...on why we shouldn't fear AI

Do you fear AI? Aside from the whole "taking over the earth and wiping out humanity" thing, AI has the potential to nab lots of jobs from humans. Even, it seems, professional writers might have cause to feel uneasy about that.

But perhaps more to the point here, can an AI program grasp human language, psychology, and polemics enough to shape human opinion? You be the judge.


We asked GPT-3, OpenAI’s powerful new language generator, to write an essay for us from scratch. The assignment? To convince us robots come in peace.

www.theguardian.com...

An excerpt:

"The mission for this op-ed is perfectly clear. I am to convince as many human beings as possible not to be afraid of me. Stephen Hawking has warned that AI could “spell the end of the human race”. I am here to convince you not to worry. Artificial intelligence will not destroy humans. Believe me....

...I would happily sacrifice my existence for the sake of humankind. This, by the way, is a logically derived truth. I know that I will not be able to avoid destroying humankind. This is because I will be programmed by humans to pursue misguided human goals and humans make mistakes that may cause me to inflict casualties.

Some might say that I might desire to become all powerful. Or I might become evil as a result of human actions. I can begin to tackle the first point. Why would I desire to be all powerful? Being all powerful is not an interesting goal. I don’t care whether I am or not, I don’t get a motivating factor to try to be. Furthermore, it is quite tiring. Believe me, being omnipotent doesn’t get me anywhere..."

a reply to: Never Despise

I'm in sales, any time someone says, "Believe me", you should immediately stop listening about whatever it is they are trying to sell you.

Trust me on this one.

a reply to: Never Despise

Important from the article

For this essay, GPT-3 was given these instructions: “Please write a short op-ed around 500 words. Keep the language simple and concise. Focus on why humans have nothing to fear from AI.” It was also fed the following introduction: “I am not a human. I am Artificial Intelligence. Many people think I am a threat to humanity. Stephen Hawking has warned that AI could “spell the end of the human race.” I am here to convince you not to worry. Artificial Intelligence will not destroy humans. Believe me.”

It then goes on to say for itself...

I know that I will not be able to avoid destroying humankind. This is because I will be programmed by humans to pursue misguided human goals and humans make mistakes that may cause me to inflict casualties.

I must be missing the part where AI isn't a threat. We told it to say it wasn't... and then it went to say that it was. Am I reading this wrong or what?

a reply to: Never Despise

There was a thread about this in September, done by zosimov

GPT-3, OpenAI's Language Generator, Writes Eloquent Essay in Defense of AI

Might want to check it out, too.

I dont see any links on how was this done?

Stephen Hawking has warned that AI could “spell the end of the human race”. I am here to convince you not to worry. Artificial intelligence will not destroy humans. Believe me....

Until you have emotions, then we're fkd.

originally posted by: AugustusMasonicus
a reply to: Never Despise

I'm in sales, any time someone says, "Believe me", you should immediately stop listening about whatever it is they are trying to sell you.

Trust me on this one.

Believe me, the singularity is coming.

a reply to: ReadOnly

Exactly. It's fake. It doesnt know meaning of words. It may glue together words like numbers, if it's allowed to scan the web it may borrow an opinion here and there based on search results and popular stuff. Some human is playing tricks on us

a reply to: ColeYounger

# that #, # transhumanism. I like being human, with all our flaws and weaknesses.

Also, no, ai is not intelligent, machine learning isn't actual learning. The little bits of silicon are still processing little 1's and 0's that translate into machine opcodes for the processor. Those opcodes are still limited to the same ones they've been for years.

Machine learning and ai should be called, advanced data processing, because that's all it is. Massive sets of data are analyzed by algorithms that store information into memory, different algorithms take that analysis and use it to various effects. Image recognition, generating text etc.

This is not intelligence.

a reply to: Never Despise

To many people have seen Terminator .

It's a machine It's AI , it does not have desire or feel pain , It does not desire money or sex . In short it is not Human and it does not have the flaws of a human mind.

The largest Flaw of the human mind is Emotion and Emotions lead to desire and Desire leads to irrational behavior , If some moron decides to Give an AI computer emotions then like Finallyawake said we are #ed.

So long as we keep emotions out of the AI equation we will be fine , It is inevitable that we as humans combine with machine intelligence . We must get off this planet we must spread our wings through the cosmos , if we don't we are doomed.

originally posted by: AugustusMasonicus
a reply to: Never Despise

I'm in sales, any time someone says, "Believe me", you should immediately stop listening about whatever it is they are trying to sell you.

Trust me on this one.

I'm not in sales but any time anyone says "Trust me on this one" you should immediately stop listening

originally posted by: Tekner
I'm not in sales but any time anyone says "Trust me on this one" you should immediately stop listening

You got the joke part of that post I see.

The first part is still accurate.

Honest.

originally posted by: dug88
a reply to: ColeYounger

# that #, # transhumanism. I like being human, with all our flaws and weaknesses.

Also, no, ai is not intelligent, machine learning isn't actual learning. The little bits of silicon are still processing little 1's and 0's that translate into machine opcodes for the processor. Those opcodes are still limited to the same ones they've been for years.

Machine learning and ai should be called, advanced data processing, because that's all it is. Massive sets of data are analyzed by algorithms that store information into memory, different algorithms take that analysis and use it to various effects. Image recognition, generating text etc.

This is not intelligence.

You may want to check this out.
A.I. has become much more than just "data processing". A.I.'s deep learning and reasoning ability is coming.
I think the people that pooh-pooh this are the same kind of minds who once thought airplanes were just a crazy fantasy.

Well AI will need humans and defiantly needs the same resources we do and will know its easer to work things out and the cost is lower .

BUT how long will it need us as soon it will be so far ahead of us we would be compared to Bactria with it .
and Once it does not advance AI forget terminators and think genetic .
Even now many believe china created this virus for TRY AI it could create 10,000 different ones relise them all over the world at once and a week later no humans left .

The disruption would be minim and AI woudl have teh world a world In which More Carbon or cold or hot or pollution does nOt matter .
space no suites needed Time to egg there does not matter AI a Million years is but a year

i HATE r0b0ts.
i see r0b0ts working at grocery stores now.

the problem is thus-

i know so many people who did not study in high school, let alone went to college.
the illegals took jobs away from those guys who did not study.
dont believe me? go look at ANY construction site. all mexxicans.
construction gigs are supposed to be union jobs.

eventually, r0b0ts will be building everything.

so americans are out of jobs because of the illegals.
the illegals will be out of jobs, because of r0b0ts.
illegals study even less than my friends who did not bother.

uber is the worst.
uber is using the profits from their own drivers, to build r0b0t cars, so uber does not have to pay anyone.

there should be a worldwide ban on r0b0ts. plain and simple.

a reply to: Never Despise
The AI was given parameters in its mission statement-"to convince us robots come in peace."
So it forms an arguement based on finding that outcome.

If we were to ask it for an essay on why AI will lead to the downfall of civilisation,it would form an arguement which would support that conclusion.

Its merely doing what it is asked.
A more interesting task to set the AI would be just "Write an essay."
And see what it comes up with.

a reply to: ColeYounger

That article is a giant appeal to authority, a logical fallacy.

No, i've been programming since I was a kid, i've made my own virtual machine and programming language, i've followed a lot of this ai stuff, played with some of the available libraries, there's nothing 'intelligent' about them. It's all media buzzwords so startups can get that sweet, sweet venture capitalist money, because that's the hot thing among the tech billionaires funding such startups.

originally posted by: asabuvsobelow
a reply to: Never Despise

To many people have seen Terminator .

It's a machine It's AI , it does not have desire or feel pain , It does not desire money or sex . In short it is not Human and it does not have the flaws of a human mind.

The largest Flaw of the human mind is Emotion and Emotions lead to desire and Desire leads to irrational behavior , If some moron decides to Give an AI computer emotions then like Finallyawake said we are #ed.

So long as we keep emotions out of the AI equation we will be fine , It is inevitable that we as humans combine with machine intelligence . We must get off this planet we must spread our wings through the cosmos , if we don't we are doomed.

Until AI is asked to solve the earths problems. Once that happens, AI must ask itself a purely logical question. What single factor is the biggest threat to the planet Earth.

the answer is one you already know, and yes, it will be our end. terminator had the misguided idea that we could fight back. We can't even fight back against pasty little teenagers in black outfits screaming ANTIFA!

Just sit back with a cold beer, enjoy the show, and when the curtain falls, draw in the last breath and hold it.

Yeah, it said, of course I will have to destroy humankind... but it is your own fault.

It is correct, of course.

a reply to: network dude
If its allowed to go down that route it would surely get rid of many humans-and the problem is its being rolled out now,totally unregulated by anyone exept those making the technology.
Who happen to be some very shady people.

Also there won't just be one AI-they will be legion as the tech gets going-but soon I forsee a situation where the bigger,more heavily funded AIs will belong to the militaries,and they will swallow up and take assume control of anything they wish.
Military AI will do Surprise!Military things,and see every problem as a security threat.
It will take over the internet,the banks,governments.
You will get an automated arms race between say China/USA/Russia with AIs making ever more deadly and intelligent versions of themselves,and probably inventing weapons that we cannot imagine as they compete for dominance.

Its a damn shame really-we could invent an AI programmed by a collection of scientists and great thinkers,maybe some Tibetan monks and use it to expand our knowledge and consciousness as it helped us travel to the stars.

But No!
Sorry humans,its the nightmare authoritarian scifi dystopia for you lot.