HOUSE_OVERSIGHT_016860.jpg

and output to sense organs and muscles, then we would reproduce, in a physical artifact, the observed manifestations of natural intelligence. Nothing observable would be missing. As an observer, I’d have no less (and no more) reason to ascribe consciousness, creativity, or evil to that artifact than I do to ascribe those properties to its natural counterparts, like other human beings.
Thus, by combining Crick’s “astonishing hypothesis” in neurobiology with powerful evidence from physics, we deduce that natural intelligence is a special case of artificial intelligence. That conclusion deserves a name, and I will call it “the astonishing corollary.”
With that, we have the answer to our three questions. Since consciousness, creativity, and evil are obvious features of natural human intelligence, they are possible features of artificial intelligence.
A hundred years ago, or even fifty, to believe the hypothesis that mind emerges from matter, and to infer our corollary that natural intelligence is a special case of artificial intelligence, would have been leaps of faith. In view of the many surrounding gaps—chasms, really—in contemporary understanding of biology and physics, they were genuinely doubtful propositions. But epochal developments in those areas have changed the picture:
In biology: A century ago, not only thought but also metabolism, heredity, and perception were deeply mysterious aspects of life that defied physical explanation. Today, of course, we have extremely rich and detailed accounts of metabolism, heredity, and many aspects of perception, from the bottom up, starting at the molecular level.
In physics: After a century of quantum physics and its application to materials, physicists have discovered, over and over, how rich and strange the behavior of matter can be. Superconductors, lasers, and many other wonders demonstrate that large assemblies of molecular units, each simple in itself, can exhibit qualitatively new, “emergent” behavior, while remaining fully obedient to the laws of physics. Chemistry, including biochemistry, is a cornucopia of emergent phenomena, all now quite firmly grounded in physics. The pioneering physicist Philip Anderson, in an essay titled “More Is Different,” offers a classic discussion of emergence. He begins by acknowledging that “the reductionist hypothesis [i.e., the completeness of physical explanations based on known interactions of simple parts] may still be a topic for controversy among philosophers, but among the great majority of active scientists I think it is accepted without question.” But he goes on to emphasize that “[t]he behavior of large and complex aggregates of elementary particles, it turns out, is not to be understood in terms of a simple extrapolation of the properties of a few particles.”15 Each new level of size and complexity supports new forms of organization, whose patterns encode information in new ways and whose behavior is best described using new concepts.
Electronic computers are a magnificent example of emergence. Here, all the cards are on the table. Engineers routinely design, from the bottom up, based on known (and quite sophisticated) physical principles, machines that process information in extremely impressive ways. Your iPhone can beat you at chess, quickly collect and deliver information about anything, and take great pictures, too. Because the process whereby computers, smartphones, and other intelligent objects are designed and manufactured is completely transparent, there can be no doubt that their wonderful

15 Science, 4 August 1972, Vol. 177, No. 4047, pp. 393-96.

57

HOUSE_OVERSIGHT_016860