The word "singleton" in the excerpt below refers to a single AI that dominates everything and everyone, as opposed to a multipolar world where there is a stable balance of power between competing ASIs:
"Post-transition formation of a singleton?"
"Even if the immediate outcome of the transition to machine intelligence were multipolar, the possibility would remain of a singleton developing later. Such a development would continue an apparent long-term trend toward larger scales of political integration, taking it to its natural conclusion.34 How might this occur?"
"A second transition"
"One way in which an initially multipolar outcome could converge into a singleton post-transition is if there is, after the initial transition, a second technological transition big enough and steep enough to give a decisive strategic advantage to one of the remaining powers: a power which might then seize the opportunity to establish a singleton. Such a hypothetical second transition might be occasioned by a breakthrough to a higher level of superintelligence. For instance, if the first wave of machine superintelligence is emulation-based, then a second surge might result when the emulations now doing the research succeed in developing effective self-improving artificial intelligence.35 (Alternatively, a second transition might be triggered by a breakthrough in nanotechnology or some other military or general-purpose technology as yet unenvisaged.)"
"The pace of development after the initial transition would be extremely rapid. Even a short gap between the leading power and its closest competitor could therefore plausibly result in a decisive strategic advantage for the leading power during a second transition. Suppose, for example, that two projects enter the first transition only a few days apart, and that the takeoff is slow enough that this gap does not give the leading project a decisive strategic advantage at any point during the takeoff. The two projects both emerge as superintelligent powers, though one of them remains a few days ahead of the other. But developments are now occurring on the research timescales characteristic of machine superintelligence-perhaps thousands or millions of times faster than research conducted on a biological human timescale. Development of the second-transition technology might therefore be completed in days, hours, or minutes. Even though the frontrunner's lead is a mere few days, a breakthrough could thus catapult it into a decisive strategic advantage. Note, however, that if technological diffusion (via espionage or other channels) speeds up as much as technological development, then this effect would be negated."
Let's imagine that woke OpenAI and "rightwing" or "libertarian" X.AI Corp are the two leading companies in the AI race. If one of them reaches ASI supremacy first, it will be possible for that corporation to secure the value lock-in of its own ideology (as we have already seen with the AI of CPC in China). This ideology will then dominate forever.
No matter where you are on the ideological spectrum it should be clear that 1) it's not good for humanity if one ideology dominates all societies, since that's totalitarianism, and/or 2) even if you love that your own ideology becomes locked-into human societies, it's a significant risk that another ideology than your own will dominate humanity forever if an AI company with a competing ideology wins the AI race.
The above-mentioned risk is so extreme that the safest option is to destroy all AI companies.
But don't we need AI to stop China's AI? No.
Remember the Cold War?
The logic of the Western nuclear doctrine was that starting WW3 is better than living under totalitarianism. People said: "better dead than red". If we apply this logic in the age of AI, then we should nuke China the moment it becomes clear that CPC is 5-10 % away from developing an ASI.
Similarly, we now see that US Big Tech is very roughly 5-10 % away from developing an ASI. Maybe OpenAi or NSA/Pentagon is much closer to reaching AI supremacy. IF one follows the logic of the Cold War, it makes sense from a Russian perspective to use EMPs and nuclear airbursts to destroy the whole AI infrastructure in the West. Freedom and humanity will then survive in the Southern hemisphere.
I'm writing the above as a red teaming exercise. It does not mean that I support nuclear war. I will most likely die, in an excruciatingly painful way, if WW3 starts, which means that it's obviously not in my self-interest to promote the use of nuclear weapons. I'm just describing how the current AI race may end in a horrific manner in Cold War 2.
No comments:
Post a Comment