Singularity Hypotheses: A Scientific and Philosophical Assessment (The Frontiers Collection) by Amnon H. Eden & James H Moor & Johnny H Soraker & Eric Steinhart

Author:Amnon H. Eden & James H Moor & Johnny H Soraker & Eric Steinhart [Eden, Amnon H. & Moor, James H & Soraker, Johnny H & Steinhart, Eric]
Language: eng
Format: epub
ISBN: 9783642325595
Amazon: 3642325599
Publisher: Springer
Published: 2013-04-03T00:00:00+00:00

most powerful real-world supercomputer. But there’s a worse problem: reflexivity.

Predicting sociological phenomena differs essentially from predicting mindless

physical phenomena. Even in a classical, causally deterministic universe, the

behaviour of mindful, reflexively self-conscious agents is frequently unpredict-

able, even in principle, from within the world owing to so-called prediction par-

adoxes (Welty 1970). When the very act of prediction causally interacts with the

predicted event, then self-defeating or self-falsifying predictions are inevitable.

Self-falsifying predictions are a mirror image of so-called self-fulfilling predic-

tions. So in common with autistic ‘‘idiot savants’’, classical AI gone rogue will be

218

D. Pearce

vulnerable to the low cunning of Machiavellian apes and the high cunning of our

transhuman descendants.

This argument (i.e. our capacity for unitary mind-simulation embedded in

unitary world-simulation) for the cognitive primacy of biological general intelli-

gence isn’t decisive. For a start, computer-aided Machiavellian humans can pro-

gram robots with ‘‘narrow’’ AI—or perhaps ‘‘train up’’ the connections and

weights of a subsymbolic connectionist architecture (Rumelhart et al. 1986)—for

their own manipulative purposes. Humans underestimate the risks of zombie

infestation at our peril. Given our profound ignorance of how conscious mind is

even possible, it’s probably safest to be agnostic over whether autonomous non-

biological robots will ever emulate human world-simulating or mind-reading

capacity in most open-field contexts, despite the scepticism expressed here. Either

way, the task of devising an ecologically valid measure of general intelligence that

can reliably, predictively and economically discriminate between disparate life-

forms is immensely challenging, not least because the intelligence test will express

the value-judgements, and species- and culture-bound conceptual scheme, of the

tester. Some biases are insidious and extraordinarily subtle: for example, the desire

systematically to measure ‘‘intelligence’’ with mind-blind IQ tests is itself a

quintessentially Asperger-ish trait. In consequence, social cognition is disregarded

altogether. What we fancifully style ‘‘IQ tests’’ is designed by people with

abnormally high AQs (Baron-Cohen 2001) as well as self-defined high IQs. Thus,

many human conceptions of (super) intelligence resemble high-functioning autism

spectrum disorder rather than a hyper-empathetic God-like Super-Mind. For

example, an AI that attempted systematically to maximise the cosmic abundance

of paperclips (Yudkowsky 2008) would be recognisably autistic rather than

incomprehensibly alien. Full-Spectrum (Super-) intelligence is certainly harder to

design or quantify scientifically than mathematical puzzle-solving ability or per-

formance in verbal memory-tests: ‘‘IQ’’. But that’s because superhuman intelli-

gence will be not just quantitatively different but also qualitatively alien (Huxley

1954) from human intelligence. To misquote Robert McNamara, cognitive sci-

entists need to stop making what is measurable important, and find ways to make

the important measurable. An idealised Full-Spectrum Superintelligence will

indeed be capable of an impartial ‘‘view from nowhere’’ or God’s-eye-view of the

multiverse (Wallace 2012), a mathematically complete Theory of Everything—as

does modern theoretical physics, in aspiration if not achievement. But in virtue of

its God’s-eye-view, Full-Spectrum Superintelligence must also be hypersocial and

supersentient: able to understand all possible first-person perspectives, the state-

space of all possible minds in other Hubble volumes, other branches of the uni-

versal wavefunction—and in other solar systems and galaxies if such beings exist

within our cosmological horizon. Idealised at least, Full-Spectrum Superintelli-

gence will be able to understand and weigh the significance of all possible modes

of experience irrespective of whether they have hitherto been recruited for

information-signalling purposes. The latter is, I think, by far the biggest intel-

lectual challenge we face as cognitive agents.