Annihilation from Within

hobbled by the fundamental limitations of lifeless machinery.12
Lifeless computers lack flexible judgment, creativity, and emotions. Emotions are an attribute of the living brain that indeed seems essential for superior intelligence. Without the diverse experience of pleasure and pain that pulsates from the living body into the brain, it has not been shown that emotions can emerge at all. Emotions impart value, and thus provide a ranking of importance to newly perceived events as well as to remembered data. By prioritizing the welter of information flowing from the brain’s senses and memory, emotions help the mind in deciding what to do. Acquired and inherited emotions stored in the amygdala and hippocampus endow the human mind with a useful image and helpful understanding of the natural and social environment. Many distinguished scholars have concluded that the full intellectual powers of the mind cannot exist in a lifeless machine.13
Social interaction among independent minds—each with its own will, self-consciousness, and emotions—is also essential for intellectual creativity. The pervasive role of language in human intelligence is proof of that. Language can lift our thoughts and sentiments above the realm of our experience. It is the magic carpet on which our mind travels beyond the outskirts of our perceptions. It enables us to form useful thoughts about new problems or new phenomena, and is indispensable for the development of science, ethics, and religion. But language lives and evolves only through social interaction. Computers, lacking emotions, cannot have social interaction and hence must operate with language that has been preprogrammed. By contrast, the human mind, when stimulated by social interaction, expands its vocabulary and idioms to capture new concepts.14
So the next challenge is evident. The lifeless computer must be coupled with the inventive capabilities and judgment of the human brain—the living brain that is both the master and slave of a human body. The scientific literature has reported hundreds of disparate research projects designed to link the brain with computers. Initially, most of these projects were conducted with modest resources, and quite a few were rather playful in design and purpose. But more recently, research projects linking the brain with computers—called brain-computer interface (BCI)—have multiplied and are now being pursued at American and European universities. These BCI projects are designed to help disabled people with prostheses to move their artificial limbs, or for those missing limbs to operate a computer or other equipment with direct signals from their brain.15 These promising projects are bound to receive increasing encouragement and financial support. Since the BCI projects serve such legitimate medical purposes and raise no ethical problems, they will proceed without provoking religious objections and risking government-imposed prohibitions on further research. And so, gradually, progress in neuroscience, biotechnology, computer sciences, and other disciplines will yield a great deal more knowledge about the triangular relationship between mind, brain, and computers.
Yet some of the best experts still seem reluctant to explore this triangular relationship as something that could be transformed into a single, integrated system. On the one hand, most studies about enhancing human intelligence remain narrowly focused on individual human beings—raising IQ one person at a time, so to speak. On the other, the votaries of Artificial Intelligence emphatically stress that they do not wish to smudge their clean work with the slithery brain—as if they feared contamination with mad cow disease. And neurologists and cognitive psychologists rarely invite computer scientists to help them design a symbiotic system combining computers and human brainpower, apart from the therapeutic brain-computer links to help disabled people.16 Such delimiting of different scientific disciplines is more common now because of the growing complexity and richness of each branch of science. Yet from time to time, bridges get built between different disciplines that bring sudden, great advances.
As night follows day, enterprising scientists will build such a bridge between computer-based Artificial Intelligence and brain science. They will begin to organize interdisciplinary projects to integrate computer systems with the mental power of living brains. Their ambitious aim will be to reach a level of intelligence well above the human range. At this time, no one can describe all the theoretical and technical problems that have to be solved for the project to be started in earnest, and it would likely require an effort with generous financing and strong support by many scientists—neither of which is available today. But at some uncertain date in the future, the search for this superhuman intelligence will become a major priority of the world’s leading countries. This will happen when it sinks in that superhuman intelligence really might be attained, and that its attainment would revolutionize all prior considerations about national security.
If U.S. intelligence organizations discovered that a nation with strong scientific capabilities—for example, China—had made significant breakthroughs on such a project, government support for a competing project in the United States would suddenly become available. Recall that America’s expensive project for the manned mission to the moon—something for which there had previously been little enthusiasm—easily garnered Congressional support when it appeared that the Soviet Union was about to accomplish the feat. Recall also that the fear Nazi Germany might acquire the atomic bomb (although in the event unwarranted) triggered the U.S. decision to launch the Manhattan Project, at the time an immense and uncertain venture. At the start of that project, none of the physicists involved could have described the full research and development program which produced the atomic bomb three years later.
A competitive race with China to build the first super-intelligent system might start sooner than most think tanks and government forecasters expect. And also sooner than I had expected when I wrote a first draft of these pages three years ago. Since then, Chinese scientists and institutes of the Chinese Academy of Sciences have published numerous articles about an “integrated” large facility for linking a brain trust and computers to work on complex policy issues. It seems this project is meant to draw on brain science and computer science (Artificial Intelligence) to combine human intelligence with high-performance computers. I am