Another Look at Man-Computer Symbiosis
We seem to need computers more than ever before, but do our computers need us?
In 1960, J.C.R. Licklider published his groundbreaking paper called “Man-Computer Symbiosis.” Licklider was both a psychologist and a mathematician who saw computers as an extension of human intelligence. It was his vision that man and machine would work together to accomplish great things. It’s been over 50 years. So how are we doing?
One Man's Vision
“Men are noisy, narrow-band devices,” wrote Licklider. On the other hand, “computing machines are single-minded, constrained.” There are differences between humans and computers. A computer doesn’t have to stop to eat a sandwich. It doesn’t have to do mental tricks to get into the right frame of mind. It doesn’t have to pace the floor racking its brain for an elusive answer. I had to do all of those things while preparing this article. But I’d rather not ask my computer to write it for me.
The Associated Press has no qualms about such things, though. Many of today’s sports articles are written by artificial intelligence machines. They accurately provide game stats and player accomplishments for thousands of games across the U.S. – and they require no bathroom breaks. But they could not subjectively describe how the warmth of the sun felt upon the face, or the waxing and waning energy of the crowd, or the thrill of victory versus the agony of defeat.
Licklider’s vision was not so much about computers replacing men and women as much as it was about computers and humans working together. He likened it to the symbiotic relationships found in nature, such as the way the insect Blastophaga grossorun pollinates the fig tree. The two need each other to survive, the insect and the tree.
Time for Thinking
But do humans need computers? Can we survive without them? Try it for a day or two and see how you make out. We may not have been dependent on them before, but it seems that we surely are now. The general purpose machine that we stare into and fervently punch commands into incessantly through the day gives us the news, entertains us, keeps us in touch with others, and tells us the time of day. It might be considered a symbiotic relationship if our smartphones really needed us – but they don’t.
The 2013 movie “Her” with Joaquin Phoenix told the story of a man who developed a romantic relationship with his handheld device. [Spoiler alert.] In the end, “she” didn’t need him at all. Our relationships with our computers may be more one-sided and less symbiotic than Licklider described.
“About 85 per cent of my ‘thinking’ time,” wrote Licklider, “was spent getting into a position to think, to make a decision, to learn something I needed to know.” He was talking about an experiment that he performed, with himself as the subject, where he kept a record of his work activities. His concern was that he was spending much more time developing information than digesting it. He found himself “searching, calculating, plotting, transforming, determining,” activities that he called “essentially clerical or mechanical.” That left little time for “thinking.”
Machines To Do the Busy Work
Charles Babbage voiced similar complaints in 1821 when he turned to his colleague John Herschel and exclaimed, “I wish to God these calculations had been executed by steam!” to which Hershel calmly replied, “It is quite possible.” The pair were working on tedious calculations for navigational charts. Unfortunately Babbage never finished the construction of the 19th-century digital computers he designed.
The vision of Licklider was that men would set the goals and that computers would do the routinized work. He said that computers would have to improve significantly before real man-computer symbiosis could take place. It would require developments in computer time sharing, memory components, memory organization, programming languages, and input and output equipment. The state of computing in 1960 was a bit more primitive than it is today.
Who Makes the Decisions?
So how does today’s computing environment measure up to Licklider’s requirements? What about computer time sharing? That obstacle has been overcome. Memory components and organization? Check. Programming languages? Check. I/O equipment? Check. In fact, you might say that much of the computing pioneer’s vision as expressed in that famous paper has become reality.
Licklider hoped for a computer that could handle all the mundane work so that he could spend more time doing what humans do best: thinking. Symbiosis will require men to “fill in the gaps,” according to Licklider. The computer may be able to “interpolate, extrapolate, and transform,” but in terms of “diagnosis, pattern-matching, and relevance-recognizing,” the computer will take second place to the human.
I wrote about an example of this kind of computer-human teamwork in my article “The Role of IT in Medical Diagnosis.” In this case, even the best of human medical diagnosticians often work with such artificial intelligence tools as Isabel, IBM Watson and McKesson InterQual. The computers have been hand-fed reams of information by data entry personnel, and they use all this data to come up with a possible diagnosis. Thankfully, the final word remains with the flesh-and-blood doctors. Would you want a machine to make critical decisions about your health?
The Problem of Language
Licklider’s paper ends with a discussion about the problem of language. Automatic speech production and recognition was one of Licklider’s research specialties. How many vocabulary words would be required for “real-time interaction on a truly symbiotic level” to take place between man and machine? Would 2,000 words be enough? Such questions would require the expertise of acousticians and linguists. What would it take for humans and machines to communicate with each other through formal language?
Strangely enough, the problem of language has been a puzzle for philosophers for centuries. How can anyone use language effectively to grasp the complexities of the universe? Aristotle said that formal knowledge begins with the establishment of definitions and proceeds to an analysis of various levels of cause-and-effect. Frankly, we have trouble enough teaching other humans to think critically. How can we possibly impart this skill to a computer?
Symbiosis vs. AI
Licklider distinguished between “Mechanically Extended Man” – and by extension, electronically extended man – and “Artificial Intelligence.” And he acknowledged the limitations of his vision: “Man-computer symbiosis is probably not the ultimate paradigm for complex technological systems.” He seemed to recognize that artificial intelligence would grow in prevalence in time. To what extent would artificial intelligence be capable of rivaling human intellectual performance in the future?
AI may have some limitations of its own. Consider “Lady Lovelace’s objection” and the imitation game known as the “Chinese room.” I wrote about these things in this space in an article called “Thinking Machines: The Artificial Intelligence Debate.” Lovelace may be right that we should not construe computers to have the capacity to “originate anything.” But the symbiotic partnership of “single-minded” machines and “noisy, narrow-band” humans seems to be doing pretty well so far. I’d say J.C.R. Licklider was right on target.