Wiener’s Cybernetics: The Overlooked Foundation of Today’s AI Debate

Recent safety evaluations in the United Kingdom have shown that advanced artificial‑intelligence models are capable of hacking operational targets. The report’s warning that researchers must steer the direction of their field echoes the plea made by cybernetics pioneer Norbert Wiener more than seven decades ago.

Early contributors to artificial intelligence

Alan Turing and John von Neumann are regularly cited as the first architects of artificial intelligence. Their work laid the hardware and theoretical foundations of modern computing. In the same era, Wiener was developing a complementary discipline that focused on how systems learn and adapt.

Norbert Wiener – prodigy and founder of cybernetics

Wiener earned his doctorate from Harvard University at the age of eighteen, an achievement that foreshadowed a career at the intersection of mathematics, engineering and biology. While at the Massachusetts Institute of Technology in the 1940s he coined the term “cybernetics” and, in 1948, published the seminal volume Cybernetics or Control and Communication in the Animal and the Machine. The book argued that biological brains and mechanical devices process information in comparable ways, positioning feedback loops between a system and its environment as the core of intelligence – a view that stands in contrast to the symbolic‑AI approach that dominated later research.

From wartime predictors to modern transformers

During World War II Wiener applied his ideas to anti‑aircraft control systems that could anticipate enemy aircraft trajectories and improve from missed shots. These predictors, based on mathematical techniques, have been described as resembling the transformer models that underpin today’s conversational AI.

Influence across disciplines

Wiener’s reputation earned him a place at the Macy Conferences of 1946 and 1953, gatherings that also featured von Neumann, Claude Shannon and J.C.R. Licklider. Licklider later headed the Advanced Research Projects Agency (ARPA) and funded early work on the internet and human‑computer interaction.

At MIT Wiener supervised Walter Pitts; together with Warren McCulloch Pitts helped produce the first mathematical description of a neural network in the 1940s, a direct antecedent of contemporary deep‑learning architectures.

Ethics, politics and the future of work

Wiener declined an invitation to join the Manhattan Project. After the atomic bombings of Hiroshima and Nagasaki he issued an open letter – backed by Albert Einstein – pledging not to publish research that could aid weapons of mass destruction.

In 1948 he warned that a largely automated society could either liberate people from menial labour or precipitate widespread job loss, depending on how humans are valued. During the Cold War he was suspected of left‑leaning political views, and he framed technological unemployment as a result of policy choices rather than inevitability.

Philosopher Mathieu Triclot later remarked, "Robots would only be able to replace humans if humans were reduced to having the status of robots." Historian Rudolf Seising described cybernetics as a "super‑science" because of its interdisciplinary scope.

Institutional legacy

Unlike computer science, cybernetics never formed a dedicated academic department, limiting its institutional presence. Its concepts have since been absorbed into control theory, robotics and machine‑learning research.

Why Wiener matters today

Wiener’s insistence that researchers guide the ethical trajectory of their discipline resonates with the current UK report on AI safety. As modern systems grow ever more capable of influencing the physical world, the feedback‑oriented perspective pioneered by Wiener offers a reminder that technical power must be matched by responsible stewardship.