Key points
- Marvin Minsky was a pioneer of artificial intelligence.
- Minsky foresaw much of how AI would ultimately evolve.
- Minsky held that thought and emotions were indelibly connected, a radical notion in psychology at the time.
Marvin Minsky of Massachusetts Institute of Technology’s Computation Center was one of the founders of the emerging field of artificial intelligence. (Vannevar Bush’s work in computing machines and Norbert Weiner’s in cybernetics had established MIT as a hub of research in the area.) While the term “artificial intelligence” was coined by John McCarthy while at Dartmouth in 1955, Minsky singularly foresaw the future of the technology. “We are on the threshold of an era that will be strongly influenced, and quite possibly dominated by intelligent problem-solving machines,” he stated in 1961, quite accurately describing where we are with AI some 65 years later.
When not thinking literally brainy thoughts (his Princeton Ph.D. dissertation was titled “Theory of Neural-analog Reinforcement Systems and its Application to the Brain-model Problem”), Minsky was able to make the foreign subject of artificial intelligence relatively understandable to a lay audience. “Scientists working on the problem of artificial intelligence are slowly learning how to exploit the structure of the modern computer,” he said, with “magnificent intellectual resources” to be released if the machines could be programmed properly.
Importantly, Minsky drew a distinction between artificial intelligence and existing “thinking machines,” the latter legitimately seen as being able to do only what they were told. Computers were only dull and unimaginative because they had been programmed to carry out tedious tasks by repeating certain designated operations. While very good at following specific instructions, the machines had no ability to solve problems that programmers had not anticipated, which is where artificial intelligence came into the picture.
“We can tell the machine that when specific instructions are lacking it should go ahead by ‘trial and error,’” Minsky explained, a giant leap in the development of the field. With artificial intelligence, “we have traded the task of step-by-step instruction for the harder but more exciting problems of supervising the exploration,” he wrote, defining programmers’ main job as to guide the search process. “The culmination of [these] efforts will come probably within our lifetimes,” Minsky predicted, while admitting that “what will happen then we cannot guess.”
While philosophers were apt to perceive humans as unique beings endowed with moral and ethical qualities that machines could never possess, those with more of a scientific background like Minsky thought otherwise. It was conceivable that computers would not only match humans in all aspects of intelligence but exceed them, he pronounced, an alarming idea at the time.
In 1966, Minsky reported in Scientific American that computers were already setting goals, making plans, generating hypotheses, and identifying relationships—a long way from simply adding or subtracting figures or playing a decent game of checkers. It was not unreasonable to believe that intuition and consciousness were next, he believed, or that people had a monopoly on wit and wisdom.
For Minsky, who was the best-known spokesperson for artificial intelligence in the 1960s, it was the discovery that thought processes could be programmed into computers that made he and others in the field so upbeat about the future of machines. Contrary to popular belief, not every step in solving a problem had to be given to computers, as they were capable of learning from past experiences.
On that basis, it was logical to conclude that the next major breakthrough in artificial intelligence would be what he called analogical reasoning—the ability to recognize similarities between two things, the key to establishing relationships. Human intelligence was currently far ahead on many levels, Minsky conceded, but machines were displaying certain subjective operations of the mind, a sign of much bigger things to come. “It is hard to say how close we are to this threshold,” he said, “but once it is crossed the world will not be the same.”
The creation of a thinking machine had always been the primary objective of artificial intelligence, and it remained so in the 1980s. There was still no consensus on what “thinking” was, however, and theories continued to float about.
Minsky remained very much in the mix, as always offering his take on the subject to both scholarly and popular audiences. In a 1983 article for Psychology Today, Minsky proposed that computers wouldn’t be able to “think” until they could be programmed to have emotions, a rather audacious proposition. Thought and emotions were indelibly connected, he held, and the field of psychology would be well served by bridging the gap between cognitive psychologists, who viewed “thinking” as the key to human behavior, and therapy-based psychologists, who focused on emotions.
On the Origin of the Healthy Mind
This was a radical notion within the psychological community, raising the question of whether such a thing could be possible. Minsky believed it could, although it would take fifty years for the computer equivalent of “emotional knowledge” to be incorporated into programs. (Doing the math, that would be 2033—now not that far away.) Computers could feel something like pride or fear, he suggested, a mechanical interpretation of the human emotion. It might take 200 years for computers to be capable of perceiving the relationship between “themselves” and the external environment, however, something that a typical child could do.
While that remains to be seen, history shows that much of what Minsky had imagined was realized.
Minsky, Marvin. (1987). The Society of Mind. New York: Simon and Schuster.
