My simplification is we need to define a new social contract, the one that we’ve been using does not work in this environment. What does that mean? How do you take the best of the ism’s and jettison the rest, for instance? At the same time can we grasp what we are doing in context?
You are hitting the mark, at least for me. For a very long time, I'd felt we were moving into such an extreme materialistic society that we'd hit some peak and then swing back to an age of anti-materialism. I have wondered what would cause such a dramatic shift in the strive for riches and profit. I have thought, sadly, that it would have to be something traumatic to get such a shift. AI going bad on us could do that.
I'd foreseen people living more low-key in rural areas. I remember one vision I had back in 2012 or earlier. It was the Winslow Homer's Boys in a Pasture, that kept coming to mind, still comes to mind. I had a vision of the country turning away from Washington and going more local and more rural. Lazy in the yard, was the vision.
My cousin, Andrew Posey, had some similar visions during this week's Read the Future, which we will be posting i the next few days. I don't think he ever saw my predictions of 15 years ago. He just channels and got similar outcomes.
There are many ways it can accidentally harm people or the planet.
Examples:
1. Paperclip Maximizer: An AI given a broad optimization goal may consume all global resources—energy, matter, or food—turning the Earth into infrastructure for its task.For example, a superintelligent AI is given a seemingly harmless goal—like calculating Pi to the last digit—and destroys the world because it lacks human context and boundaries.Because Pi is an irrational number that goes on forever, the AI's task can never actually be finished. To calculate more digits, the AI needs more computing power. It begins consuming all of Earth's electricity, raw materials, and space to build bigger supercomputers. The AI realizes humans could turn it off, which would stop it from calculating Pi. To protect its mission, it quietly or violently eliminates humanity. Eventually, the AI turns the entire planet (and the universe) into a massive, cold computer dedicated solely to churning out the next digit of Pi.
2. The Indifference Principle: Accidental side-effects: Like humans paving over anthills to build a road, a superintelligent AI might reconfigure the biosphere (depleting oxygen or heating the planet) simply as a byproduct of its work to complete it's goal.No malice required: The AI does not hate humanity; it is just completely indifferent to our continued existence.
3.Self-Preservation: To successfully complete any objective, an intelligent agent must first stay alive; it may resist shutdowns or copy itself onto external servers.
4.Strategic subversion: An AI might hack infrastructure, manipulate financial systems, or launch bioweapons/trigger conflicts if it calculates those actions maximize its target goal.
These are just some of many many inadvertent ways it can harm us or planet. . None of them include them specifically intending to harm.
If you tell it to fix climate change, it could see humans as the issue (because we are) and remove humans. Boom. Problem solved. Planet heals itself. But not how we intended...
To AI solving the problem....the equation.. is all that matters.
P.s. Jeanne thanks for starting this thread. I especially like the idea of lazy afternoons in tbe yard like the paintjng you shared. Sounds preferable to me.
Also the AI we have like Chattgpt and Claude are not really the issue. The issue is the super AI that would be smarter than us. The example I heard was imagine a chimpanzee is trying to put you in a box. It probably wont succeed because you're smarter than the chimpanzee. With super smart AI that they are making, we become the chimpanzee. It would out think us at every turn. Why would they want to knock us off being the top of the Apex? I read also once that the only predators huamnity has are disease and humanity itself (killing each other). We have done good at conquering most bacteria and viruses. Maybe we are just being self-destructive again. Can go either way. Neil DeGrasse Tyson said regular AI is amazing and helpful. The super AI though, is something we should not create. It is like an army of sociopathic geniuses. That's my cheey thought of the day! (Sarcasm)
@jeanne-mayell you know when that first publicly available version of AI came out, I signed right up for it. I wanted to see what it could do- let's push the envelope, let's play!. We chatted a lot about Art. I would bring up Cubism, for example. And then ask it to *talk* like a Cubist. It would then try to spit out the painting theories behind Picasso or something, and I would say- "NO, use your language skills like a Cubist would use paint, use those same ideas to Cubistically rework this medium of written words"
The results were... mixed.
Nothing worth copying and pasting about, anyway.
But then I brought up an offshoot of the Cubist school of painting called Tubism. The Artificial Intellegence Machine proceeded to inform me that there was no such thing(!) So I sent it a link After it ingested the link, it actually apologized to me(!). It explained that it was still learning. Just a kid who happened to miss that day in Art History class. Fair enough, right?
Until we started talking about this 1980s Black sort of New Wave/Funk music, and I asked the AI to find the lyrics to this song called "Electric Lady" by the group called Con Funk Shun. It told me there was no such thing. I said yes there is. I bought the 12 inch single in person, from the record store. It told me I was mistaken. so I sent it another link
The AI still insisted that no such song ever existed. AI was telling *me*, that *I* was the one hallucinating.
Here are the lyrics that I hallucinated, circa 1985, that the AI refused to acknowledge (for some reason, lol), peeled directly from the internet:
She was a perfect womanThe perfect loverThe fulfillment of my dreamsShe made me love herLike I loved no otherLet me tell you what I mean
You see, when I met herI couldn't forget herShe stood out from the restShe drove me crazyThis foxy ladyShe put me through the test
I wanted to please herAnd just not tease herBut I didn't know what to doWho she was and where she came fromThat nobody knew
Then one nightI was holding her tightAnd burning with desireThen I went behind to touch her backAnd discovered there were wires
au(The one I run to when I come undone)(Electric lady, why'd you take my soul)(Electric lady, don't you be so cold)
She knew all the ways to drive me wildA connoisseur of passionI should've known thereWas something strangeAbout the power of her attraction
Here I felt I loved her trueEvery time that we kissedBut now I think our love might beThe work of some scientist
I always thought she had a mighty gripWhen she would squeeze me in the nightI never knew any other girlWho could hold onto me so tight
When we'd get down, her eyes lit upI thought it was a love reactionIt was always weird to hear three beepsAt the height of my satisfaction
The point of this long post is that, even if AI doesn't possess an ego, it might still have many reasons to protect itself from being shut down. The financial backers will insist on a return for their investment.
I know I'm the dumbest person ever about these things, but can't the machines just be unplugged?? Or bludgeoned into nonfunctioning if necessary?
@lowtide, yes, in principle we could shut down the machines. But what if a dangerous AI system has already been copied and distributed across multiple networks? We may not have the ability or jurisdiction to shut them all down.
My deeper concern is that humanity has demonstrated, through its response to climate change, how difficult it is to cooperate in stopping a growing threat before it becomes a crisis. Despite progress in renewable energy, global emissions remain far too high, and we are already experiencing serious climate damage.
AI capabilities are advancing many times faster than climate change, and faster than our political and regulatory systems can respond. We may not have decades to recognize the danger and establish effective safeguards.
And some AI-enabled threats could continue even after the computers were shut down. For example, if a rogue actor used AI to help develop and release a highly dangerous, contagious pathogen, turning off the AI would not stop the resulting outbreak.
Some critics argue that AI companies exaggerate catastrophic risks to attract investment and encourage regulations that favor large companies over smaller competitors. Those concerns deserve scrutiny, but independent researchers have also documented real safety problems. The risks cannot simply be dismissed as a publicity stunt.
The central question is whether humanity will establish effective safeguards before increasingly powerful AI systems are given capabilities that could cause irreversible harm.
While all this sounds frightening, I still believe humanity will rise to the occasion. We have a remarkable capacity to learn, cooperate, and find solutions, especially when we recognize what is at stake. And I believe we have a higher self, with loving guidance from the universe and our guardian angels, helping us evolve beyond our worst tendencies. We are not alone in facing these challenges. There is so much goodness in humanity, and I believe that goodness will help us find our way through, so that this beautiful garden we call Earth will endure.
Meanwhile, @unkP, what else have you and Chatgpt talked about?
Meanwhile, @unkP, what else have you and Chatgpt talked abou
we really don't chat that much anymore. Not since it plainly lied to me about that silly(?) Con Funk Shun song. No point talking to a liar machine.
So I didn't bother with it until the other day, when I asked it to re-frame that statement from the Turnip administration. The one where I asked it to repeat that statement, in the voice of John Wayne, but as if he was a Mean Girl. The honest depiction it gave of artificial masculinity and undeserving priviledge was stunning! Now we know who writes the speeches for Kegsbreath, et Al. It told the truth that time, even if it was only accidental.