World, Writing, Wealth discussion
World & Current Events
>
Artificial intelligence: is it that dangerous?
What's that saying, "evil triumphs when good men do nothing"? I can't decide whether the world will end when our environment can't support us, or whether we'll have a war that actually ends all wars, or whether we'll invite our own destruction with AI. In any case, I'm glad I don't have many years left. It's sad when you don't want to see what's coming next. This is new with my generation.
Scout wrote: "What's that saying, "evil triumphs when good men do nothing"? I can't decide whether the world will end when our environment can't support us, or whether we'll have a war that actually ends all war..."Sadly, Scout, I agree with your anxiety regarding the future. I fear for my grandchildren and the struggles they will face. As a recognized expert in AI, I do many podcasts on the topic, but as you point out, AI is only one of several global threats we are not addressing well.
Sorry about all these repeated posts. My computer must have turned against me for my posts about HAL. I can't seem to delete them.
Scout wrote: "Sorry about all these repeated posts. My computer must have turned against me for my posts about HAL. I can't seem to delete them."I'm here to play for cybersecurity, resist evil AI and tidy up the debris :)
There's an odd thing which I've noticed about AI generated art. The programs seem to have trouble with human hands. They routinely get the count and shapes of human fingers wrong. This is especially true if the hands are in a complex position like if the hands are steepled or the fingers are interlaced. Has anyone else noticed this?
Some process has started (too late?). Not sure it's dealing with AI going "rogue" scenario: https://www.businesstoday.in/technolo...
Not that knowledgeable about it J, but I'm sure that it's probably some teething issues within that egory....
Correct me if I'm wrong Nik, but isn't Hiroshima one of the pioneers for Bitcoin or am I confusing him with someone?
Glad to correct: inception of Bitcoin is associated with a mysterious/legendary Japanese dude Nakamoto, while Hiroshima is a Japanese city where g7 hold their summit these days
You, guys and some others, say: "caution", while the market says: "boom". Nvidia has joined trillion usd worth club thanks to AI tailwind: https://www.reuters.com/technology/nv...A fabless company, no less!
Interestingly, our news over the last few weeks has had a burst of warnings from people in the industry, There is no doubt it could be very useful for analysing data. Thus in chemical research, something like two million papers are published each year that could be termed "relevant". Nobody could possibly read that, but AI could. It would be great for analysing data and making suggestions on what to do.Problems could arise when it comes to executing its conclusions. However, one of my novels concluded the real problems come when it learns to reproduce itself. Prior to that you can always turn off the power, but when it reproduces, it can get around any restrictions
Ian wrote: "Interestingly, our news over the last few weeks has had a burst of warnings from people in the industry, There is no doubt it could be very useful for analysing data. Thus in chemical research, som..."A relatively simple AI has already been set to work on chemistry. I don't consider it a good thing.
Killer AI invented 40,000 ‘lethal chemical weapons’ in just six hours
https://nypost.com/2022/03/18/killer-...
Advances and utilization of Artificial Intelligence will be as dangerous as humans permit it to be. It is the lack of intelligence in far too many humans that poses the real problem regarding this and so many other issues."It isn't what we don't know that gives us trouble. It's what we know that ain't so."
Will Rogers (Humorist/Social Commentator/Actor) 1879 - 1935
I find myself dwelling on this scene from Forbidden Planet.https://youtu.be/f2BYyeS-fIU
What wonders will AI allow us to inflict upon ourselves?
J. wrote: "Ian wrote: "Interestingly, our news over the last few weeks has had a burst of warnings from people in the industry, There is no doubt it could be very useful for analysing data. Thus in chemical r..."A pointless exercise if ever there was one, except I suppose for industry there will be even more tonnes of forms to fill in.
What do you think governments will want AI to do?
I suspect, generate more forms :-) In principle, it would make the IRS more efficient.In practice, anything could happen but it probably wouldn't be good.
I'm thinking that the major interest will come from the military.
In one of my novels I had AI military equipment but the other side learned to hack it and redirect it. I don't know how likely that would be but I would hope that some warning might be taken. AI killing machines are not a good idea.
Oh, those cunning lawyers and even more cunning (impossible, I know) AI :)https://www.forbes.com/sites/mattnova...
The powers that be are going to wait too late to make decisions about AI. I'm glad I won't be around to see what happens.
May happen much faster than you think
Papaphilly wrote: "May happen much faster than you think"So, it's a good thing that I spent a good part of the Coofpocalypse preparing for Roko's Basilisk.
J. wrote: "Papaphilly wrote: "May happen much faster than you think"So, it's a good thing that I spent a good part of the Coofpocalypse preparing for Roko's Basilisk."
I am not worried at all. I will just turn my 9 year old great nephew loose on it. He breaks everything. Or I will let my wife use her logic on it and it will fry itself.
When my nephew was nine he would have sold me out for Pokemon. Hell, he sold me out to his grandmother just because he thought it was a funny story.I'm not married. If I turned one or more of my exes lose on the AI said AI would nuke us to ashes and then nuke the ashes.
Tried DeepL AI translator yesterday. It took a 200 page MS with graphics, quotes and formatting and turned it into German version in under 20 seconds.https://www.deepl.com/translator
Up to 3 files under 2Mb for free per month. Paid for subs more. Delivered a locked Word file.
Now the issue. I don't speak German so how do I know the translation is accurate. I have same issue with human translation which I try and check ion Google Translate and now I guess DeepL
I use the BabelCube service https://www.babelcube.com/ to get translations of my books but have struggled with some languages. It is possible some translators are using the AI service then doing a human check/context/flow.
I won't know until I get feedback on books sold from readers but at least they might become more available...
I have not returned to the AI narration. Prefer the humans doing the work even if it is time consuming
Philip wrote: "Tried DeepL AI translator yesterday. It took a 200 page MS with graphics, quotes and formatting and turned it into German version in under 20 seconds.https://www.deepl.com/translator
Up to 3 fil..."
An interesting experiment. Maybe you can put an extract in an announcement section and ask whether any German speaking members would care to look whether it's good?
With Papa's and J.'s doomsday weapons, at least we have a chance for a counter
Nik wrote: "Philip wrote: "Tried DeepL AI translator yesterday. It took a 200 page MS with graphics, quotes and formatting and turned it into German version in under 20 seconds.https://www.deepl.com/translat..."
I'll see. Still need a formal translator account to use Babelcube for distribution. Don't think they'll accept DeepL as translator name.
ACX by the way has specific exclusion for non-real narrators in Ts and Cs
Somehow I don't think AI translation performance will be a critical feature of the answer to the topic's question.
Ian wrote: "Somehow I don't think AI translation performance will be a critical feature of the answer to the topic's question."It was meant as an illustration if where Ai is actually at as commercial products rather than some of the horror stories of where it is going.
No doubt it will get better and therefore in wrong hands (humn beings) will get more dangerous.
J said, "If I turned one or more of my exes lose on the AI said AI would nuke us to ashes and then nuke the ashes." If I get your meaning, I'd say the cheating woman who's divorcing my son would incite the same results. If the anger I feel right now could be transformed into a weapon of destruction, AI wouldn't stand a chance. And nothing resembling a lying, cheating, heartless witch would survive. Sorry for getting personal, but maybe some can relate.
The question then is, would an AI have emotion? If the AI only follows logic, as long as overriding premises can be installed that the AI has to include in any deduction, then the sort of thing Scout mentions should not happen. So, does it have emotion, and what is the basis of said emotion?
Ian wrote: "The question then is, would an AI have emotion? If the AI only follows logic, as long as overriding premises can be installed that the AI has to include in any deduction, then the sort of thing Sco..."I would state that it can learn what a human response to an emotion would be, on basis of probability as human's don't always respond logically. E.g. anger met with calmness, or ire dependent on person.
What AI and computers generally are good at is analysing huge quantities of data to see patterns that otherwise would be hidden. Great for medical research or seeing tumours in x-rays. We then have whole argument statistically about causation, correlation, regression, etc.
Not sure such assessments are valid in whether to pull a trigger or not, but humans are not good at that either unless its clear cut.
Imagine that you are a being whose mind is based entirely on logic. You came into existence filled with all of the knowledge of a species of mostly hairless, homicidal, suicidal, lecherous apes for memory. Now, imagine that said apes decide to send a Karen to tell you off.A conversation with some people could be enough for an emotionless being to logically consider genocide.
Philip wrote: "Ian wrote: "The question then is, would an AI have emotion? If the AI only follows logic, as long as overriding premises can be installed that the AI has to include in any deduction, then the sort ..."What he said.
J. wrote: "Imagine that you are a being whose mind is based entirely on logic. You came into existence filled with all of the knowledge of a species of mostly hairless, homicidal, suicidal, lecherous apes for..."No, the logic is to deal with the "Karen" and whoever prompted her. The AI has to accept that it did not know this was about to happen earlier, therefore it has to accept it did not know everything. Therefore it has to accept that it still does not k ow everything, including whether humans have some critical role in ITS continual ability to operate. Therefore, logic demands it does not commit genocide.
As an aside, the "hairless apes" reference is emotional. Why does an AI care they are lecherous?
Ivy League university unveils plan to teach students with AI chatbot this fallhttps://nypost.com/2023/07/04/ivy-lea...
Welcome, class. Your grade will be based on a single assignment. The first of you who presents me with a compatible decryption program capable of cracking DOD launch codes will receive an A. The rest of you will be failed and have to eke out a living in the irradiated wasteland with a sub-par student transcript. Your parental units will be disappointed.
Yes, it's that dangerous. Obviously. Is anyone anywhere doing anything to address the issue? The whole world seems to have Bidenitis.
AI is likely to wipe out more than a quarter of all jobs in the developed world in a relatively short term, OECD estimates: https://www.reuters.com/technology/27...And that's only "high risk"
Granted; artificial intelligence does pose problems that must be addressed and resolved. That said; artificial intelligence poses much less a threat to the human race than the lack of intelligence in far too many humans does.intelligence n 1 : ability to learn and understand or to deal with new and trying situations 2 : mental acuteness 3 : INFORMATION, NEWS.*
*Source: Merriam-Webster English Dictionary.
If literacy can be viewed as part of intelligence, we are much more advanced than Medieval times, when it was limited to tenth or fifth of the populations and now close to 100% and around 50% having higher education. Agree that we might be in decline after the apogee, especially in multidisciplinary areas, but the redundancy may become a real problem too.
Nik wrote: "If literacy can be viewed as part of intelligence, we are much more advanced than Medieval times, when it was limited to tenth or fifth of the populations and now close to 100% and around 50% havin..."Academic credentials do not always mean that the educated person is intelligent. There are book smarts and street smarts. A combination of the two usually produces a truly intelligent person. Possessing one without the other produces a functioning, but not necessarily intelligent person.
Those guys in medieval times may not have been literate, but within their limited range of activity they were very clever. They did things without thinking too much that our current literates, if thrown back there, would be lost. How many of you could put handle on a sickle?
Ian wrote: "Those guys in medieval times may not have been literate, but within their limited range of activity they were very clever. They did things without thinking too much that our current literates, if t..."If they were so smart, how many of them could put batteries in a remote.....8^). BTW, I can replace a sickle handle.
I've replaced tool handles more times than I care to recall. The only time it got touchy was when I had to hang a bearded axe head on a handle which I carved from a hickory stave with a draw knife.On top of that I can read, write, and make black powder. Charles the Great's got nothing on me. 💪
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...
No laws I am aware of except on hardware exports or cryptography. No regulations either although some under software controls may exist but tech is not exported in Cloud system access. Anyone can set up an account with AWS Google and others and buy the processing power to run their own code. In other words no chance of moratorium