Didn’t see a comment section there, but it’s a bit sad that the AI worshipers apparently missed all the Terminator movies.
Labeling Musk and Hawking as Luddites is ridiculous, I suspect it’s really just publicity-seeking. Musk is spending a billion to research AI, seems to me that’s as far as you can get from Luddism.
The problem is not malevolence, it’s amorality coupled with the unintended consequence, coupled with the speed of the machine. Similar to the “grey-goo” scenario for nano-tech replicators. To get a feel for what the latter would look like (times a few gazillion) check out this video:
Domain restrictions will be necessary. Just like messing around with genetically modified human pathogenic viruses in non-P4 labs.
People don’t tend to worry about C3PO, whose influence stops at his robotic appendages, R2D2 is a little more problematic inside a Death Star and forget about HAL…
There’s no reason to think that technological breakthroughs will be always in the interest of the future of the human race.
= Ivy Mike.
Really, does a strong AI really present that much more threat than an amoral natural intelligence in the same situation? Consider what communists and jihadis have done, and continue to attempt.
Communists and Jihadis don’t think / plan / calculate / react at the speed of light.
You can neutralize a jihadi by splattering his brains. An AI’s brains are theoretically embedded in your computer, your smartphone, your networked thermostat, your car, the traffic signals, the emergency dispatch system, the power company’s nuclear station…
Yep. The HAL scenario…
I do not consider jihadists amoral. They are about as far from it as you can get, since their motivations are moral imperatives derived by their interpretation of religious teachings.
To get a feel for the amoral you have to consider the context in which decisions are made. For example, some AI has been demonstrably shown in the past to suffer from a “horizon effect”, i.e. a path chosen for action, although optimal in the short run turned out not to be so in the long run. Why? Because the AI has to deal with a practical barrier upon how much time it can take to process the data presented to it and yet maintain some semblance of real-time interaction. Along similar lines, it not a big stretch to see similar possibilities when context is not included. For example, imagine an AI running a power grid. No one thought to impose a moral value upon say airport ground control. If the AI treats all members of the grid equally without this extra consideration, an airport ground controller could find him or her self at a suddenly blank radar screen because the AI was trying to optimize power distribution over a city as the greater good. This is a trivial example. The REAL issue will be of course “the devil” lost in the detail. Although it won’t be “The Devil” per-se, just a lack of consideration. Or part of the “you can’t think of everything” scenario. Some would argue that this is a window in time, that eventually AI can catch up and eventually surpass us in moral epistemology. Asimov believed that it was at least possible. Some think that the odds are against us and that we won’t survive through the window. Others (Vinge, Gibson) have written fiction along the lines of the hope that human intelligence or consciousness reaches “the singularity” at least as closely as AI might in order survive. We shall see. To paraphrase A.C. Clarke, “The truth is stranger than fiction”…
I’d be honored to be in the company of Musk and Hawking! To be honest, though, I’m not as worried about strong AI as I am about the oligarchs in charge of Amazon, Apple, Facebook, Google, and Microsoft using their soft AIs to take away the last vestiges of our freedom.
Consider cars: to work, they require gasoline, oil, brake fluid, and so forth. No mere accident could enable a car to forage in the wild and refuel from tree sap: this would demand engineering genius and hard work. – Eric K. Drexler
Fear mongering is a popular sport. Elon Musk does it daily – if it’s not AI, it’s climate change or “ULA spies”. This award is apt.
Didn’t see a comment section there, but it’s a bit sad that the AI worshipers apparently missed all the Terminator movies.
Labeling Musk and Hawking as Luddites is ridiculous, I suspect it’s really just publicity-seeking. Musk is spending a billion to research AI, seems to me that’s as far as you can get from Luddism.
The problem is not malevolence, it’s amorality coupled with the unintended consequence, coupled with the speed of the machine. Similar to the “grey-goo” scenario for nano-tech replicators. To get a feel for what the latter would look like (times a few gazillion) check out this video:
https://www.youtube.com/watch?v=6lv-sVNHIX0
Domain restrictions will be necessary. Just like messing around with genetically modified human pathogenic viruses in non-P4 labs.
People don’t tend to worry about C3PO, whose influence stops at his robotic appendages, R2D2 is a little more problematic inside a Death Star and forget about HAL…
There’s no reason to think that technological breakthroughs will be always in the interest of the future of the human race.
= Ivy Mike.
Really, does a strong AI really present that much more threat than an amoral natural intelligence in the same situation? Consider what communists and jihadis have done, and continue to attempt.
Communists and Jihadis don’t think / plan / calculate / react at the speed of light.
You can neutralize a jihadi by splattering his brains. An AI’s brains are theoretically embedded in your computer, your smartphone, your networked thermostat, your car, the traffic signals, the emergency dispatch system, the power company’s nuclear station…
Yep. The HAL scenario…
I do not consider jihadists amoral. They are about as far from it as you can get, since their motivations are moral imperatives derived by their interpretation of religious teachings.
To get a feel for the amoral you have to consider the context in which decisions are made. For example, some AI has been demonstrably shown in the past to suffer from a “horizon effect”, i.e. a path chosen for action, although optimal in the short run turned out not to be so in the long run. Why? Because the AI has to deal with a practical barrier upon how much time it can take to process the data presented to it and yet maintain some semblance of real-time interaction. Along similar lines, it not a big stretch to see similar possibilities when context is not included. For example, imagine an AI running a power grid. No one thought to impose a moral value upon say airport ground control. If the AI treats all members of the grid equally without this extra consideration, an airport ground controller could find him or her self at a suddenly blank radar screen because the AI was trying to optimize power distribution over a city as the greater good. This is a trivial example. The REAL issue will be of course “the devil” lost in the detail. Although it won’t be “The Devil” per-se, just a lack of consideration. Or part of the “you can’t think of everything” scenario. Some would argue that this is a window in time, that eventually AI can catch up and eventually surpass us in moral epistemology. Asimov believed that it was at least possible. Some think that the odds are against us and that we won’t survive through the window. Others (Vinge, Gibson) have written fiction along the lines of the hope that human intelligence or consciousness reaches “the singularity” at least as closely as AI might in order survive. We shall see. To paraphrase A.C. Clarke, “The truth is stranger than fiction”…
I’d be honored to be in the company of Musk and Hawking! To be honest, though, I’m not as worried about strong AI as I am about the oligarchs in charge of Amazon, Apple, Facebook, Google, and Microsoft using their soft AIs to take away the last vestiges of our freedom.
Fear mongering is a popular sport. Elon Musk does it daily – if it’s not AI, it’s climate change or “ULA spies”. This award is apt.