Skynet Anyone?

[stromboli]

http://motherboard.vice.com/read/a-global-arms-race-to-create-a-superintelligent-ai-is-looming

​Forget about superintelligent AIs being created by a company, university, or a rogue programmer with Einstein-like IQ. Hollywood and its AI-themed movies like Transcendence and Her have misled the public. The launch of the first truly autonomous, self-aware artificial intelligenceâ€"one that has the potential to become far smarter than human beingsâ€"is a matter of the highest national and global security. Its creation could change the landscape of international politics in a matter of weeksâ€"maybe even days, depending on how fast the intelligence learns to upgrade itself, hack and rewrite the world's best codes, and utilize weaponry.

In the last year, a chorus of leading technology exp​erts, like Elon Musk, Stephen Hawking, and Bill Gates, have chimed in on the dangers regarding the creation of AI. The idea of a superintelligence on Planet Earth dwarfing the capacity of our own brains is daunting. Will this creation like its creators? Will it embrace human morals? Will it become religious? Will it be peaceful or warlike? The questions are innumerable and the answers are all debatable, but one thing is for sure from a national security perspective: If it's smarter than us, we want it to be on our sideâ€"the human race's side.

Now take that one step further, and I'm certain another theme regarding AI is just about to emergeâ€"one bound with nationalistic fervor and patriotism. Politicians and military commanders around the world will want this superintelligent machine-mind for their countries and defensive forces. And they'll want it exclusively. Using AI's potential power and might for national security strategy is more than obviousâ€"it's essential to retain leadership in the future world. Inevitably, a worldwide AI arms race is set to begin.

As the 2016 US Presidential candidate for the Transhumanist Party, I don't mind going out on a limb and saying the obvious: I also want AI to belong exclusively to America. Of course, I would hope to share the nonmilitary benefits and wisdom of a superintelligence with the world, as America has done for much of the last century with its groundbreaking innovation and technology. But can you imagine for a moment if AI was developed and launched in, let's say, North Korea, or Iran, or increasingly authoritarian Russia? What if another national power told that superintelligence to break all the secret codes and classified material that America's CIA and NSA use for national security? What if this superintelligence was told to hack into the mainframe computers tied to nuclear warheads, drones, and other dangerous weaponry? What if that superintelligence was told to override all traffic lights, power grids, and water treatment plants in Europe? Or Asia? Or everywhere in the world except for its own country? The possible danger is overwhelming.

Given the AI Imperative, there's really only two likely courses of action for the world

Below is something simple I've designed that's tautological in nature called the "AI Imperative." It demonstrates why an AI arms race is likely in humanity's future:

1) According to experts, a superintelligent AI is likely possible to create, and with enough resources, could be developed in a short amount of time (such as in 10-20 years).

2) Assuming we can control this superintelligent AI, whoever launches it first will likely always have the strongest superintelligence indefinitely, since that AI can be programmed to undermine and control all other AIsâ€"if it allows any others to develop at all. Being first is everything in the superintelligent AI creation game (imagine if you were first to develop the Atomic bomb, and then also had the power to limit who else could ever develop one).

3) Whichever government launches and controls a superintelligent AI first will almost certainly end up the most powerful nation in the world because of it.

Given the AI Imperative, there's really only two likely courses of action for the world, even though there's four major possibilities on how to proceed. The first is to make AI development illegal all around the worldâ€"similar to chemical weapon development. However, people and companies probably would not go for it. We are a capitalistic civilization and the humanitarian benefits of AI are too promising to not create it. Stopping development of technology has never really worked, either. Someone else just ends up eventually doing itâ€"either openly or in secretâ€"if there's gain or profit to be made.

The other option is to be the first to create the superintelligent AI. That's the one my money is onâ€"the one America is going to pick, regardless which political party is in office. America's military will likely spend as much of its resources as it needs to make sure it has exclusivity or majority control in the launch of a superintelligent AI. I'm guessing that trillions of dollars will be spent on AI development by the American military over the next ten years, regardless of national debt, economic conditions, or public disagreement. I'm betting that engineers, coders, and even hackers will become the new face of the American military, too. Our new warriors will be geeks working around the clock in the highest security environment possible. Think the Manhattan Project, but many more times in size and complexity.

Of course a third option is that AI is developed via a broad international consortium. However, nuclear weapon proliferation shows why, at least so far, this idea will likely not come to passâ€"at least on a worldwide level. As long as powerful nations like Russia and China independently push their flavor of social policy, economic development, and government operations (many of which largely mirror their leader's desires), this is unlikely to work or be accepted. This is because we're not talking about good old fashioned teamwork exploring outer space together on the space station or stopping third-world civil wars and genocides, as the respected United Nations sometimes is involved in. We're talking about military power and protection of our families, citizenry, and livelihoods. There's much less room for cooperation when it concerns such personal matters.

A fourth option, one that I believe may be inevitable in the long run, is that all nations unite democratically and politically under one flag, one elected leadership, and one government, in an effort to better control the technology that is ushering in the transhumanist ageâ€"such as superintelligent AI. Then, all together, we create this intelligence. I like the sound of this from a philosophical and humanitarian point of view. The problem with it is such a plan takes time and many proud people to swallow their egos and cultural differencesâ€"and with only about 10 to 20 years before superintellitent AI is created, no one is going to push hard for that option.

So, inevitably, we are back to our looming dog-eat-dog AI arms race. It may not be one filled with nuclear fallout shelters like yesteryear, but it will show all the signs of the most powerful nations and the best minds they posses vying against one another for an all-important future national security. More importantly, it's a winner-takes-all scenario. The competition of the century is set to begin.

Scary shit. I for one do not think that political forces in opposition would find common ground- just like the nuclear arms race, up to the point of competing AI capabilities and a static state not unlike what we have now.

I've read a few stories about such a concept with varying outcomes. But the concern is we could see some cyber attacks that dwarf anything previously seen; complete bankruptcy of monetary systems, total infrastructure shutdowns, interruption of services at an unprecedented scale. And don't even think of the potential weapons possibilities.

Might have to go back to my survivalist roots.

[Hydra009]

Scary shit. I for one do not think that political forces in opposition would find common ground- just like the nuclear arms race, up to the point of competing AI capabilities and a static state not unlike what we have now.

I've read a few stories about such a concept with varying outcomes. But the concern is we could see some cyber attacks that dwarf anything previously seen; complete bankruptcy of monetary systems, total infrastructure shutdowns, interruption of services at an unprecedented scale. And don't even think of the potential weapons possibilities.

Might have to go back to my survivalist roots.

https://www.youtube.com/watch?v=elqY0cuXTo8

Heh.

But seriously, don't pack that bug-out bag quite yet.  There's still a long way to go before strong AI is a reality.  I joke that machines are already smarter than humans, but that's only true in some *ahem* special cases.

And thankfully, the nations capable of building a strong AI - tech-savy and rich nations - are almost exclusively Western.  So, it's unlikely that nations hostile to the West would develop strong AI.  And it's a pretty safe bet that AI researchers share your concerns and would introduce strong AI to friendly AI before it ever sees the light of day.  I'm a big fan of the scaffolding approach, rather than going from calculator to godhood.

Also, who knows, maybe a strong but friendly AI would be a real asset in resolving persistent global problems and heading off future problems.  We already know that we're in for a serious challenges up ahead.  This could be a (figurative) godsend.

[Gawdzilla Sama]

I had a very crazy friend disappear on August 27th, 1997. He recently emerged from a secret hidey-hole in Montana.

[Atheon]

I, for one, welcome our new robot overlords.

[_Xenu_]

I wouldn't worry too much about a strong AI emerging soon, even the most enthusiastic predictors consider it decades off. When it does arrive though, it will be the end of history one way or another. Either all of our problems will be solved, or humanity will be destroyed.

[stromboli]

I wouldn't worry too much about a strong AI emerging soon, even the most enthusiastic predictors consider it decades off. When it does arrive though, it will be the end of history one way or another. Either all of our problems will be solved, or humanity will be destroyed.

It took something like 6 years to go from theory to actual product in the construction of the A bomb. there was more than one country making the effort, but obviously we were the most committed and spent much more to arrive at the end result. Several scientists from other countries, some of them Jewish, came here to escape persecution from other governments and ideologies. At that point we had a huge edge in technology and intellectual resources.

I see similarities in this with the A bomb. Now we have China as the foremost contender, rather than Germany. Russia basically stole the A bomb, so in that sense the worry is to keep them from repeating the effort. But China, being a legitimate super power and capable of applying as vast a set of resources as we have, could certainly make the effort. We do not have the huge edge of technology or brain power we had then, in fact our attitudes toward science are now being throttled to some extent by the backward nature of our politics.

I don't think it would take decades, based on the effort made. Our technological capabilities are vastly greater than in the 40's. Speed of information gathering is approaching quantum levels. Although AI is a huge leap to be sure, it is nonetheless attainable. We held absolute supremacy for a few short years with the A bomb. Whoever develops AI would hold an even greater supremacy, operating across a vast network that is literally worldwide and continually improving. The author's contention that it could ultimately be benevolent to me flies in the face of past events. I think the potential outcome could be disastrous. Let us hope it isn't.

And you are right. Either all our problems will be solved or humanity will be destroyed.

[Hydra009]

Whoever develops AI would hold an even greater supremacy, operating across a vast network that is literally worldwide and continually improving.

Psht.  I wish it was actually worldwide.  Partially worldwide atm.

[trdsf]

AI is to computer technology as fusion power is to physics -- it's always 20 to 50 years in the future, regardless of what year you're in.

Alternatively, AI was accidentally created in 1976, and the machine was smart enough to let on, and has been infecting every computer made ever since, just to annoy humanity.

More seriously -- the unspoken assumption is that what's being worked on is not AI -- artificial intelligence -- but AHLI -- artificial human like intelligence.  For our purposes, not any intelligence would do.

I'm not sure where I stand on the possibility of AI.  Certainly extremely 'clever' machines exist, but as of right now, I don't think there's any system that there's even a consensus that it exhibits actual intelligence.  Right now, I'm leaning towards the Douglas Hofstadter view that intelligence is an abstraction layer that is both fundamentally dependent upon, and simultaneously fundamentally unaware of, its hardware layer -- I mean, if you had to explicitly direct the firing of each neuron in order to have a thought, you'd starve to death before you could even register the thought that you were a little hungry.

So I'm inclined to think that a programmatic approach is probably not going to lead to a system that can be legitimately called intelligent.  A system may well pass the Turing Test, but the Turing Test is a measure of language parsing ability, not comprehension.  You can submit a natural language question on the Google search line and it will respond appropriately the vast majority of times, but in no way can you say that the Google search engine genuinely has any knowledge or understanding of your question.

Or, to go back to the first two sentences of this post: the second one is (probably) a joke, the first one is (probably) not.

[stromboli]

In any case, time will tell. Personally I'm glad because now I can ramp up the paranoia and convince the wife we need to buy land in Colorado. But living in a cave might be a bit problematic.....

[Hydra009]

In any case, time will tell. Personally I'm glad because now I can ramp up the paranoia and convince the wife we need to buy land in Colorado. But living in a cave might be a bit problematic.....

You could always come after a few days, claim an angel told you all sort of stuff, carve out a large empire, and achieve reverence akin to godhood centuries after your death.

[stromboli]

You could always come after a few days, claim an angel told you all sort of stuff, carve out a large empire, and achieve reverence akin to godhood centuries after your death.

Lol. I look good in my hiking shorts, rock sandals,  my camo vest and tool belt.

[trdsf]

Psht.  I wish it was actually worldwide.  Partially worldwide atm.

I want to know what that little dot in the middle of Greenland is...

[stromboli]

Nice to know one can get cash in the middle of the Sahara.

[_Xenu_]

As this topic seems to be dwindling into irrelevancy, I thought I would post some serious reading on the matter. This article argues that AI will be our downfall, and addresses common arguments that say otherwise. It also links to more reading on the subject.

http://philosophicaldisquisitions.blogspot.se/2014/07/bostrom-on-superintelligence-4.html

[Gawdzilla Sama]

I want to know what that little dot in the middle of Greenland is...

Weather station.

Seriously. No Such Agency told me it was a weather station.