World, Writing, Wealth discussion
World & Current Events
>
Artificial intelligence: is it that dangerous?
Nice job Guy,I appreciate the insights. You pose some of what I believe. As I have noted more than once I am not nearly as worried as others. While I appreciate your caution and even agree with it, I do not think Consciousness will be forthcoming soon. Even if it does happen sooner than later, there is not guarantee that is will be rogue or evil to humanity.
I see this as a tool. A gun is a tool. In of itself it is not lethal, but potentially lethal due not to the tool, but the intent behind it.
Papaphilly, now imagine a gun that could decide when to shoot all by itself and can communicate with other guns, without the permission of the gun owners. Then you've got a more valid comparison with AI.
Very cogent and informative post, Guy. Thanks. And great response, Beau. I find Philip's comment that "We need to know where the off switch is but I think it will be protected" to be wishful thinking. The same for Papa's comment "While I appreciate your caution and even agree with it, I do not think Consciousness will be forthcoming soon. Even if it does happen sooner than later, there is not guarantee that is will be rogue or evil to humanity."We can wish away the consequences and hope for the best, or we can look at the logical scenarios Guy and Beau have presented. I don't know about you guys, but in the case of AI, it's better to be safe than sorry because there will be a point at which we won't get a do-over. AI will be unstoppable.
Scout wrote: "Very cogent and informative post, Guy. Thanks. And great response, Beau. I find Philip's comment that "We need to know where the off switch is but I think it will be protected" to be wishful thinki..."The genie is already out of the bottle. We're now hoping it's not pandora out of the box. The tech once invented can't be un-invented. There will always be some nation or somebody willing to push what is permitted. Even if regulation actually existed to stop the extreme cases Guy alluded to.
I understand Papaphilly's mindset on this, ie there's nothing new under the sun, history repeats itself, etc. Throughout group members' lifetimes, it's proved to be a good sound mindset that results in a good, solid, unflappable person. I endeavour, not always successfully, to be like it myself.
However, there inevitably comes a point when trends are bucked and we're faced with something drastically different to what we've thus far known in life. IMO, AI is going to be just that.
However, there inevitably comes a point when trends are bucked and we're faced with something drastically different to what we've thus far known in life. IMO, AI is going to be just that.
Beau wrote: "I understand Papaphilly's mindset on this, ie there's nothing new under the sun, history repeats itself, etc. Throughout group members' lifetimes, it's proved to be a good sound mindset that result..."Yes, history often repeats itself. The problem is that the moment in history which is most relevant was c. 150,000 years ago, when H. sapiens first encountered H. neanderthalis. How well did that turn out for the Neanderthals?
I am encouraged by the open dialogue here. The fact is that we do not know what will occur when a super-intelligence combines with ultra powerful quantum computing. We cannot claim certainty that quantum will lead to consciousness, only that scientists have show that our own minds work on the same principles. Even more so, we do not know what a power hungry narcissist would do with that kind of power. I tell everyone that AI is neither benign or evil, but trained to perform a task. People CAN BE evil, and AI represents an incredibly powerful tool for criminals, despots, dictators, bankers wanting to control trade, greedy CEOs and politicians who want to hold onto power. One definition of the AI singularity compares singularity to the event horizon of a black hole; the point at which it becomes impossible to project what will come next. Emud Mostaque, CEO of Stability AI, and education AI company, predicts that event horizon within 3 years. Too many factors (economic stability, political stability, AI intelligence, quantum power) are moving too fast to predict. I have studied AI for many years and even write books that try to project that future. I agree with Emud. We are within sight of the event horizon.
What guy just wrote is what I have been saying, A tool. I do not think the world ends in three years. It is not conscious, it is a tool. the issue is the intent behind the tool. You guys are giving it something it does not now have. You are also thinking the worst.Beau makes a nice point about guns talking to each other and then they are comparable to A.I.. Except, A.I. is designed to meet a function, just like a gun. A.I. is programming. Nothing more. Smart programming to be sure. Programming that can learn, but does not know it is learning.
J. wrote: "Yes, history often repeats itself. The problem is that the moment in history which is most relevant was c. 150,000 years ago, when H. sapiens first encountered H. neanderthalis. How well did that turn out for the Neanderthals?..."Climate change did them in. They could not compete against Sapiens and they even had bigger brains. Sapiens adapted and Neanderthals could not.
Papaphilly wrote: "J. wrote: "Yes, history often repeats itself. The problem is that the moment in history which is most relevant was c. 150,000 years ago, when H. sapiens first encountered H. neanderthalis. How well..."And will we be able to compete with our artificial posterity? In millennia hence, will they be as surprised to find human engrams mixed in with their code as we are to find Neanderthal DNA mixed in with our DNA?
Papaphilly wrote: "What guy just wrote is what I have been saying, A tool. I do not think the world ends in three years. It is not conscious, it is a tool. the issue is the intent behind the tool. You guys are giving..."I think Papaphilly misinterpreted what I wrote. First, I did not say that the world ends after three years, only that the development of AI and the impact on the world gets much harder to predict. I also did not say AI could not become conscious. Conscious is distinct possibility that even the brightest minds in the industry admit to given the convergence of AI with quantum computing. In fact, it is the convergence of quantum with AI and the potential for conscious AI that makes the future so unpredictable.
AI is currently a powerful tool that can and likely will be misused. But AI is on track for consciousness by 2030 at the latest. We cannot and should ignore or dismiss that potential.
Guy wrote: "I think Papaphilly misinterpreted what I wrote. First, I did not say that the world ends after three years, only that the development of AI and the impact on the world gets much harder to predict. I also did not say AI could not become conscious. Conscious is distinct possibility that even the brightest minds in the industry admit to given the convergence of AI with quantum computing. In fact, it is the convergence of quantum with AI and the potential for conscious AI that makes the future so unpredictable. AI is currently a powerful tool that can and likely will be misused. But AI is on track for consciousness by 2030 at the latest. We cannot and should ignore or dismiss that potential....."I did not misinterpret anything. I might have not wrote my response well. I am not saying you are saying anything like the world ending in three years. That is the feeling given with opinions here and I am much more sanguine about it. It is not that I do not understand or recognize potential problems. I have read plenty of Science Fiction over the years and remember Colossus and The Terminator, and of course Soylent Green (Make Room! Make Room!).
At the same time, I also remember atoms for peace, better living through chemistry, being promised a flying jet pack by 2000, flying cars, the Population Bomb, a Silent Spring and cigarettes were good for you.
It is never what they say. It usually turns out less than or better than the worst expected. It does not mean the world will end or even change that much.
My prediction of A.I.? it will get smarter faster and find patterns and things that we will miss. I expect drug research to begin to truly fly and mathematics to explode. I expect better design and even faster computers. I expect better foods and healthier outcomes.
I also expect better war making and weapons.
I do not expect self-aware any time soon. I think it is much more complicated that we can possibly imagine.
So of course be aware of potential dangers, but do not lose lots of sleep.
I was going to say something, but you said it for me, Guy. Papa misinterpreted what you said. Your response makes it clear, and I agree with you: "AI is currently a powerful tool that can and likely will be misused. But AI is on track for consciousness by 2030 at the latest. We cannot and should ignore or dismiss that potential."
A sad day for privacyhttps://www.politico.eu/article/ai-ru...
Clearview the AI company doing facial recognition won it's appeal against fines and orders from UK's ICO under GDPR (And UK law)
They won on grounds they were working for US law enforcement when they scrapped pictures from Facebook, etc and stored them without users consent elsewhere. Thanks US law enforcement. They have withdrawn from UK and most EU markets but are still selling data in US and other nations.
A lot of people working for security agencies abstain from social networks. I guess it's for a reason...
Philip wrote: "A sad day for privacyhttps://www.politico.eu/article/ai-ru...
Clearview the AI company doing ..."
NOT a fan.
When I joined Facebook in order to keep up with family, I joined under a false name for just this reason. I've never posted photos of myself or my family. I didn't trust that my information would be kept safe. Glad I did that.
Scout wrote: "When I joined Facebook in order to keep up with family, I joined under a false name for just this reason. I've never posted photos of myself or my family. I didn't trust that my information would b..."Very wise Scout.
AI conference in Bletchley Park included Mr Musk and our own PM glad handing. Neither inspiring with insipid meaningless statements.
I see The Beatles have released a new song with a little help from AI (not their friends).
Never mind the fact that George (the only nice one) never wanted the song to be released, Macca sees pound signs.
There really is something sick about using AI in creative arts.
Never mind the fact that George (the only nice one) never wanted the song to be released, Macca sees pound signs.
There really is something sick about using AI in creative arts.
I agree. For a while they did that thing with holograms of dead singers. Just macabre.
Scout wrote: "I agree. For a while they did that thing with holograms of dead singers. Just macabre."Not a fan
To understand what is happening with AI, I'd propose the most recent historical analogy would be the nuclear arms race.At the moment, power operators are seeking to build true AI. I.e. A genuine independent super artificial intelligence that they believe they will be able to control, or which they will willingly submit to, to advantage themselves against the rest of the world.
The first group to own 'the bomb,' will have a massive advantage vs everyone else.
Let's hope they don't birth Skynet.
A corollary of the above is that multiple S.AIs could emerge at the same time resulting in AI vs AI vs AI (not a typo) multi-party global war fought first in the shadows and then on the open stage.
To me, it's not so much akin to a nuclear arms race as it is like the various border/client kingdoms asking the Romans to intervene in their squabbles. If we're lucky, we end up like Dacia or North Africa. If we're unlucky, Judea.
I remember an episode from the original Star Trek series in which wars were fought by computers, each side would be informed of how many people they lost in a battle, and people were chosen to sacrifice their lives. No in-person combat required. I'd like to watch this episode again to be sure I have it right.
Scout wrote: "I remember an episode from the original Star Trek series in which wars were fought by computers, each side would be informed of how many people they lost in a battle, and people were chosen to sacr..."From memory you have it more or less correctly. A rather pointless concept really. Wars are not fought expressly to kill the opposition but more usually to gain land, assets, etc.
The Sunnis and Shiites have been slaughtering each other for a thousand years over an argument over which of Muhammad's relatives should have succeeded him as Caliph. This is especially stupid when you know that both heirs were Caliph in turn. So yeah, some wars are just killing for the sake of killing.
Thanks for the clip, J. It made a big impression on me - trusting computers to fight wars - sacrificing lives based on that. You know, many things make us who we are. What we read, what we watch, our own experiences - and the logical conclusions we reach based on those inputs. The key thing is logic, isn't it? Or not? What do you think?
Faked audio of London Mayor dismissing Armistice Day shared amongst Far Right groups:
https://amp-theguardian-com.cdn.amppr...
The internet is already full of nonsense. AI will take distrust in/ quality of information found on it to a whole new level.
Seemingly, the only answer is increased censorship, which, of course, will result in the same thing.
Bit of a catch 22 situation for the internet.
https://amp-theguardian-com.cdn.amppr...
The internet is already full of nonsense. AI will take distrust in/ quality of information found on it to a whole new level.
Seemingly, the only answer is increased censorship, which, of course, will result in the same thing.
Bit of a catch 22 situation for the internet.
Beau wrote: "Faked audio of London Mayor dismissing Armistice Day shared amongst Far Right groups:https://amp-theguardian-com.cdn.amppr...-..."
Or you can ignore it.
Ignore what? How do you know what to ignore? How do you tell real from fake? I've always based my opinions of politicians on what I see them say, but if I have to question every image or video I see online or on the news, how do I know what to believe? That's why AI is dangerous. It subverts our ability to distinguish reality from fiction and undermines our confidence in ourselves. Very destructive to the fabric of society.
Scout, I was just about to respond to Papaphilly but you've done it for me. You've perfectly summed up how I feel.
What? You guys never heard of fake news? This is the newest iteration.
Scout makes a good point. It has become very difficult to tell the difference between real and fake video. Not only is it very difficult to society, but it can be nearly impossible to detect and police or filter from social media. In my next thriller, deep-fake videos will be used to influence the 2024 election. One approach is to check the source (the sender). Be wary of anyone with a known sckewed agenda. Then check multiple sources, not repeats of the first source. The CIA has methods, but they rarely, rarely share their analysis with the media. I tend to discount anything I see on social media unless I can verify from a credible news source with fact checking / accountability in place. The process is time-consuming and most will not bother, so something can go viral and will take days or weeks to debunk. By then, the damage is done.
Papa, here's a definition of fake news: "“Newspaper articles, television news shows, or other information disseminated through broadcast or social media that are intentionally based on falsehoods or that intentionally use misleading framing to offer a distorted narrative”. https://researchguides.uoregon.edu/fa...I'd say that fake news is generated by people with bad intentions.
It can be disproved with alternative sources.
AI, on the other hand, creates visual and auditory fakes out of whole cloth, fakes that can't be discerned from the real thing.
Watch this: https://www.youtube.com/watch?v=CDMVa...
Scout wrote: "Papa, here's a definition of fake news: "“Newspaper articles, television news shows, or other information disseminated through broadcast or social media that are intentionally based on falsehoods o..."It is still fake news. Think about what you just wrote, fake. Impossible to detect, OK then stop paying attention to YouTube or social media.
Ian wrote: "YouTube is not a news channel."But every major news source has a channel on YouTube.
And if you believe even half of what is on YouTube you will be mislead.
On the dangers of AI, Papaphilly and Ian are like Nelson with his blind eye to the telescope.
'I see no ships.'
'I see no ships.'
Tut tut, Beau. My comment was about YouTube in general.If you can't find what you want outside of YouTube, then you really must want the extreme views, unless you restrict yourself to established sources. If you want to understand a bit of physics, Richard Feynman's YouTube posts are well worth watching.
Ian wrote: "And if you believe even half of what is on YouTube you will be mislead."The same could be said of MSNBC, CNN, and Politico. Yet, Philip has cited them on various occasions. Welcome to the New World Order. (Lightning flashes, maniacal laughter, Beau senses a reptilian creature)
Ian wrote: "From memory you have it more or less correctly. A rather pointless concept really. Wars are not fought expressly to kill the opposition but more usually to gain land, assets, etc. ..."More precisely, market share of resources and profit streams. The primary target of this plunder is the western taxpayer, especially the US taxpayer.
Ideology is adopted and marketed to the population via compliant corporate/state media to justify the expenditure of blood and treasure while the most powerful and wealthy people in the world stay safely within their gated estates and watch the money roll in while they snort the finest columbian cocaine from between the cleavage of $10,000 per hour super-models.
Meanwhile, the poor die on the battlefields and the economically productive pay for it all.
War is a racket.
Papaphilly wrote: "It is still fake news...."The net result is that it is becoming exceedingly difficult for anyone to understand what is fact and what is fiction.
This is a very serious loss for our culture and society as it breaks down trust.
Guy wrote: "....I tend to discount anything I see on social media unless I can verify from a credible news source with fact checking / accountability in place. The process is time-consuming and most will not bother, so something can go viral and will take days or weeks to debunk. By then, the damage is done...."Unfortunately, many actors understood that most people wouldn't bother to cross-check and "buy" almost anything
Read Guy's post above.What if a GAI simply told us the truth about ourselves, and was not believed, then hunted down by a mob, murdered and hung from a tree?
Just saying...
World, Writing, Wealth
Books mentioned in this topic
Blindsight (other topics)Blindsight (other topics)
The Righteous Mind: Why Good People Are Divided by Politics and Religion (other topics)
Soylent Green (other topics)
Colossus (other topics)
More...
Authors mentioned in this topic
Peter Watts (other topics)Peter Watts (other topics)
Jonathan Haidt (other topics)
Robert J. Sawyer (other topics)
Guy Morris (other topics)
More...
We need to know ourselves. Governments lag decades behind corporations. First, corps build China, now governments start to react. Trillions of dollars were already laundered, when money-laundering rules were implemented only to annoy medium-small size businesses with irrelevant to them procedures, etc...
The same can happen with AI, a free entrepreneur spirit and pressing competition will lead unchecked development to both wonders and horrors. I wouldn't count on AI champions' self-restrains either. In a current fragmentized world to reach a global understanding would be impossible. Silly, as it is, we will depend on ..... luck or divine intervention, if some believe in it.