EigoPro 1級 読解 — The Ghost in the Machine: On the Limits of Knowing Artificial Minds

For decades, the prospect of artificial consciousness has hovered at the margins of respectable science, dismissed as the province of speculation rather than inquiry. Yet as machines grow eloquent enough to discuss their own apparent inner lives, the question has acquired an uncomfortable urgency. The deepest obstacle, however, is not engineering but philosophy. Even if we could replicate every function of a sentient brain, it would remain an open question whether subjective experience—what philosophers call the "hard problem"—accompanies the performance at all. To build a machine that behaves as though it feels is one achievement; to know that it actually feels is another, and the gulf between them may be unbridgeable. The difficulty is that behavioural evidence, however dazzling, is constitutively silent on the matter. A system might answer our every question about pain, longing, or wonder with exquisite nuance, and this would be perfectly compatible with the system's being devoid of any inner life whatsoever. Eloquence is not awareness. No conceivable behavioural test can settle the question, because the very thing in dispute—the presence of felt experience—leaves no signature in conduct that mere mechanism could not in principle counterfeit. We are confronted not with a shortage of data but with the irrelevance of data to the thing we most wish to know. Some respond that we face the same predicament with one another and proceed regardless. True enough: I cannot inspect your experience either. But here the parallel falters. When I attribute a mind to a fellow human, my inference rests on a presumed similarity of bodily constitution—a shared biology that an artificial system, however articulate, does not possess. The argument from analogy, precarious even among humans, grows more precarious still when extended to silicon. We are thus left with what might be called the verification gap: a permanent, constitutive impossibility of confirming the very property whose presence would most profoundly alter our moral universe. A machine could one day persuade us utterly, and the question of whether anyone was home behind the persuasion might never, in principle, be answered at all.

新規登録 または ログイン が必要です