The Real Risk Of Artificial Intelligence
ii. Disabilities are objective empirical matters. Likewise, what computational architecture and surgeries are deployed by means of a brain or a computer (what computationalism takes to be necessary), and exactly what physical and chemical processes underlie (what mind-brain individuality theory takes to become essential), are objective empirical questions. All these are questions to be settled by appeals to proof reachable, in principle, to any qualified audience. According to such objections, irrespective of how apparently intelligently a computer acts, and no matter what mechanics and underlying physical processes allow it to do so, it might nevertheless be disqualified from truly be unintelligent as a result of its lack of abstract qualities essential for true intelligence. These presumed qualities are, in principle, introspectively discernible to the subject with them and nobody else: they're "personal" adventures, as it is sometimes put, to which the subject has "privileged access."
Is one of the most frequently heard objections to AI and a recurrent theme in its literary and cinematic portrayal? Whereas we have strong inclinations to state computers see, hunt, and infer items we've scant inclinations to state they ache or itch or encounter ennui. But to be sustained, this objection requires reason to think that thought is inseparable from feeling. Perhaps computers are only dispassionate thinkers. Indeed, far from being considered indispensable to logical thought, fire traditionally has been presumed antithetical to it. Alternately -- if emotions are crucial to enabling overall human level intelligence -- maybe machines might be artificially endowed with them: or even with subjective qualia (below) at least using their functional equivalents.
For their mathematical and other apparently high-level intellectual skills have no feelings or feelings... so, what they do – nevertheless "high-level" -- isn't real thinking.
Required for overall human-level intelligence -- this, conversely, would reply: Which iii. The Lack of Feelings Objection 2. Objection II: It Thoughts are specific biological brain processes. Whatever else it could take for general human-level intellect -- besides has proven more difficult than might have been expected. Languages are symbol systems and (sequential structure) computers are symbol crunching machines, each with its own proprietary instruction set (machine code) in which it contrasts or compiles instructions couched in high level programming languages such as LISP and C. High-level computer languages express imperatives which the system "knows" procedurally by translation to its native (and similarly critical) machine code: their constructions are basically instructions. Natural languages, on the other hand, have -- perhaps principally -- declarative functions: their constructions include descriptions whose comprehension seems fundamentally to need rightly relating them to their referents in the world. Furthermore, high level computer language instructions have unique machine code compilations (for a specified server), whereas, the same all-natural language constructions may bear unique meanings in various linguistic and extralinguistic contexts. Contrast "the child is from the pen" and "the ink is in the pen" in which the first "pencil" must be understood to imply a kind of enclosure along with the second "pen" a sort of writing implement. Commonsense, at a word, is how we understand this; however would a machine understand, unless we can somehow endow machines? While the holy grail of complete all-natural language understanding remains a distant dream, here as elsewhere in AI, piecemeal progress has been made and locating application in grammar checkers; data retrieval and information extraction systems; natural language interfaces for matches, search engines, and question-answering systems; and even limited machine translation (MT).
Towards the "scaled up" and interconnected human-level capabilities Nothing that it is like, subjectively, for a pc. The "light" of consciousness isn't on, inwardly, for them. There's "no 1 home." This is a result of their lack of sensed qualia. To equip computers with sensors to detect environmental conditions, for example, would not thereby endow them with all the personal sensations (of warmth, cold, color, pitch, and so forth) that accompany sense-perception in as such personal sensations are what consciousness is made of.
Whether this outcome will spell defeat for the strong AI thesis which Objection: Picture Holding that thoughts essentially are biological brain procedures -- yields yet another debate:
is subject to dispute. Scalability problems seem grave enough to scotch short term optimism: never, on the other hand, is a long time. If Gödel unlimited mathematical skills, or rule-free flexibility, or feelings, are needed, maybe these could be artificially produced. Gödel aside, feeling and flexibility clearly seem related in us and, equally clearly, much manifest stupidity in computers is tied to their own rule-bound inflexibility. However, even if overall human-level intelligent behavior is unnaturally unachievable, no blanket indictment of AI threatens obviously from this whatsoever. Rather than conclude from such a generality that non-human AI and piecemeal high-level AI aren't real intelligence, it might possibly be better to conclude that noninvasive AI (such as intelligence in lower life-forms) and piecemeal high-level skills (such as those of human "idiot savants") are real intelligence, albeit piecemeal and non.
CommentairesSuivre le flux RSS des commentaires
Ajouter un commentaire