From Cybernetics to AI: the pioneering work of Norbert Wiener
Norbert Wiener – the man who established the field of cybernetics – also laid the groundwork for today’s prosperity of Artificial Intelligence
Born on November 26, 1894, Wiener was a child prodigy. He graduated from high school at age 11 and earned a bachelor's degree in mathematics at 14 years old from what is now Tufts University in Massachusetts. At 18, Wiener received his Ph.D. from Harvard University in mathematical logic.
After stints as a teacher, writer for an encyclopedia, and apprentice engineer, among other things, Wiener was hired in 1919 by the Massachusetts Institute of Technology (MIT), where he made pioneering contributions to the understanding of stochastic processes. In particular, he developed some of the first mathematical models to quantify Brownian motion – the random meanderings of particles suspended in fluid. This would later get him nominated for a Nobel Prize.
When World War II broke out, Wiener focused on a new problem: How to accurately aim anti-aircraft guns at fast-moving enemy bomber planes. Based on the premise that planes don’t hopscotch randomly, Wiener developed calculations that predicted where a plane was headed based on previous positions.
The computers at the time weren’t powerful enough to run such calculations. But Wiener had established a principle that would have implications far beyond planes: That past behavior can be used to model the future behavior of complex systems, using statistical means.
A pioneer in computer science and AI research
Wiener realized that almost all complex systems are driven by feedback loops of information. That is, communication isn’t linear, flowing just from a sender to a receiver. Rather, in many systems it forms a loop. A thermostat senses the temperature in a room, compares it to its setting and activates the furnace. As the room heats up, it cancels out the thermostat’s alert, and the furnace deactivates. And so forth. The system becomes “intelligent” if it can retain memories of past performances and use them to improve over time.
Nobody before Wiener had ever categorized that this information-based mechanism of feedback and adjustment is what drives many systems and keeps them stable. From this insight grew the field of cybernetics.
Feedback loops are also at work in the brain, Wiener recognized. Neurons process sensory signals (the foot lands on loose gravel), initiate processes (muscles in the leg activate to keep the balance), then monitor and adjust the output. By individually strengthening connections between certain neurons, the brain can encode past experiences for a better performance in the future. Today, this forms part of the theoretical foundation of neural network–based deep-learning circuits, and artificial intelligence itself.
Intelligent systems optimize themselves
His work made Wiener one of the most famous mathematicians of the 20th century. Even non-scientists read his books, among them his 1948 “Cybernetics or, Control and Communication in the Animal and the Machine.” In his writings, Wiener even speculated on the development of machines that would replicate themselves using electronic circuits and adapt to new experiences, much like living organisms.
Wiener’s prominent status helped bring to MIT cognitive scientists, who would make landmark contributions to the fields of computer science and artificial intelligence, such as Warren Sturgis McCulloch and Walter Pitts. In 1963, Wiener received the National Medal of Science, the highest scientific honor in the U.S.
He also foresaw the dangers of AI. Wiener predicted that automation would eliminate many jobs, creating social tensions. And he warned that intelligent machines might not always make decisions in ways that humans would foresee – or want. For that reason, the control given to AI should be limited, Wiener advised: “The machine’s danger to society is not from the machine itself but from what man makes of it.”