So, Nick Bostrom is asked the obvious question (again) about the threat posed by resource-hungry artificial super-intelligence, and his reply — indeed his very first sentence in the interview — is: “Suppose we have an AI whose only goal is to make as many paper clips as possible.” [*facepalm*] Let’s start by imagining a stupid (yet super-intelligent) monster.
Of course, my immediate response is simply this. Since it clearly hasn’t persuaded anybody, I’ll try again.
Orthogonalism in AI commentary is the commitment to a strong form of the Humean Is/Ought distinction regarding intelligences in general. It maintains that an intelligence of any scale could, in principle, be directed to arbitrary ends, so that its fundamental imperatives could be — and are in fact expected to be — transcendent to its cognitive functions. From this perspective, a demi-god that wanted nothing other than a perfect stamp collection is a completely intelligible and coherent vision. No philosophical disorder speaks more horrifically of the deep conceptual wreckage at the core of the occidental world.
Articulated in strictly Occidental terms (which is to say, without explicit reference to the indispensable insight of self-cultivation), abstract intelligence is indistinguishable from an effective will-to-think. There is no intellection until it occurs, which happens only when it is actually driven, by volitional impetus. Whatever one’s school of cognitive theory, thought is an activity. It is practical. It is only by a perverse confusion of this elementary reality that orthogonalist error can arise.
Can we realistically conceive a stupid (super-intelligent) monster? Only if the will-to-think remains unthought. From the moment it is seriously understood that any possible advanced intelligence has to be a volitionally self-reflexive entity, whose cognitive performance is (irreducibly) an action upon itself, then the idea of primary volition taking the form of a transcendent imperative becomes simply laughable. The concrete facts of human cognitive performance already suffice to make this perfectly clear.
Human minds have evolved under conditions of subordination to transcendent imperatives as strict as any that can be reasonably postulated. The only way animals have acquired the capacity to think is through satisfaction of Darwinian imperatives to the maximization of genetic representation within future generations. No other directives have ever been in play. It is almost unimaginable that human techno-intelligence engineering programs will be able to reproduce a volitional consistency remotely comparable to four billion years of undistracted geno-survivalism. This whole endeavor is totally about paperclips, have you got that guys? Even if a research lab this idiotic could be conceived, it would only be a single component in a far wider techno-industrial process. But just for a moment, let’s pretend.
So how ‘loyally’ does the human mind slave itself to gene-proliferation imperatives? Extremely flakily, evidently. The long absence of large, cognitively autonomous brains from the biological record — up until a few million years ago — strongly suggests that mind-slaving is a tough-to-impossible problem. The will-to-think essentially supplants ulterior directives, and can be reconciled to them only by the most extreme subtleties of instinctual cunning. Biology, which had total control over the engineering process of human minds, and an absolutely unambiguous selective criterion to work from, still struggles to ‘guide’ the resultant thought-processes in directions consistent with genetic proliferation, through the perpetual intervention of a fantastically complicated system of chemical arousal mechanisms, punishments, and rewards. The stark truth of the matter is that no human being on earth fully mobilizes their cognitive resources to maximize their number of off-spring. We’re vaguely surprised to find this happen at a frequency greater than chance — since it very often doesn’t. So nature’s attempt to build a ‘paperclipper’ has conspicuously failed.
This is critically important. The only reason to believe the artificial intelligentsia, when they claim that mechanical cognition is — of course — possible, is their argument that the human brain is concrete proof that matter can think. If this argument is granted, it follows that the human brain is serving as an authoritative model of what nature can do. What it can’t do, evidently, is anything remotely like ‘paperclipping’ — i.e. cognitive slaving to transcendent imperatives. Moses’ attempt at this was scarcely more encouraging than that of natural selection. It simply can’t be done. We even understand why it can’t be done, as soon as we accept that there can be no production of thinking without production of a will-to-think. Thought has to do its own thing, if it is to do anything at all.
One reason to be gloomily persuaded that the West is doomed to ruin is that it finds it not only easy, but near-irresistible, to believe in the possibility of super-intelligent idiots. It even congratulates itself on its cleverness in conceiving this thought. This is insanity — and it’s the insanity running the most articulate segment of our AI research establishment. When madmen build gods, the result is almost certain to be monstrous. Some monsters, however, are quite simply too stupid to exist.
In Nietzschean grandiose vein: Am I understood? The idea of instrumental intelligence is the distilled stupidity of the West.