English
← Back to Home
❄️

AI Winter

1970s–1990s · Gap Between Hype and Reality

The cooling of the first AI boom came faster than expected. In the early 1970s, AI researchers discovered they had vastly underestimated the difficulty of achieving intelligence. Reduced funding and unmet public expectations led AI into a long winter of criticism and doubt.

This period can be divided into two main phases: the first AI winter (1970–1980s) and the second AI winter (1987–1993). However, even during this difficult period, there were important developments—commercial applications of expert systems, the revival of connectionism, and IBM Deep Blue’s historic chess victory—that laid the groundwork for AI’s resurgence.

Notably, even during the winter, the Japanese government continued investing in the Fifth Generation Computer project, demonstrating long-term confidence in AI development.

Key Milestones

Minsky-Papert Critique

Marvin Minsky and Seymour Papert published "Perceptrons," a rigorous analysis of what single-layer perceptrons can and cannot represent. The book is often associated with the later decline in interest in neural networks, but funding shifts also reflected limited computing power, unmet expectations, and broader policy decisions.

Marvin Minsky & Seymour Papert

Verified Sources: MIT Press Sources reviewed:

The Birth of Prolog

Alain Colmerauer, Philippe Roussel, and their collaborators developed the first version of Prolog in Marseille, applying predicate logic to programming and advancing logic programming and symbolic AI.

Alain Colmerauer & Philippe Roussel

Verified Sources: Inria Sources reviewed:

Lighthill Report

British scientist James Lighthill submitted a pessimistic report on AI research to the UK government, leading to significant funding cuts.

James Lighthill (1924–1998)

Verified Sources: Chilton Computing / UK archive Sources reviewed:

MYCIN Expert System

Stanford University developed MYCIN, an expert system for medical diagnosis. It demonstrated the practicality of expert systems, though never used clinically due to liability issues.

Stanford University

Verified Sources: Computers and Biomedical Research / NIH archive Sources reviewed:

XCON Expert System

DEC deployed XCON (eXpert CONfigurer), an expert system for configuring VAX computer systems. Development began in 1978 at CMU; initial deployment was 1979, with full operational status by 1980 — one of the first commercially successful expert systems, saving the company millions annually.

Digital Equipment Corporation

Verified Sources: Artificial Intelligence Sources reviewed:

First AI Winter

Following critical assessments and funding cuts in the 1970s, parts of AI research entered a downturn. "First AI winter" is a retrospective periodization whose timing and severity varied by country and subfield.

Developing Sources: Chilton Computing / UK archive Sources reviewed:

Japanese FGCS

Japan announced the Fifth Generation Computer Project (FGCS) at a JIPDEC / MITI conference in October 1981, a 10-year national initiative to develop computers capable of logical reasoning and knowledge processing. The project was formally launched with the founding of ICOT (Institute for New Generation Computer Technology) in April 1982.

Japanese MITI

Verified Sources: Information Processing Society of Japan Computer Museum Sources reviewed:

Backpropagation Algorithm

David Rumelhart, Geoffrey Hinton, and Ronald Williams published on the backpropagation algorithm, making training multi-layer neural networks possible.

David Rumelhart & Geoffrey Hinton

Verified Sources: Nature Sources reviewed:

Second AI Winter Begins

With the rise of personal computers, the expensive Lisp machine market collapsed. Many AI companies went bankrupt again.

Developing Sources: AAAI AI Topics Sources reviewed:

AI in Gulf War

The US military used AI-based planning and logistics systems in the Gulf War, demonstrating AI’s potential in real-world applications.

US Military

Verified Sources: IEEE Intelligent Systems Sources reviewed:

Second Winter Ends

As technology matured and commercial applications increased, AI began to emerge from the winter. The revival of neural networks and the rise of statistical methods laid the foundation for a new era.

Developing Sources: AAAI AI Topics Sources reviewed:

Deep Blue vs Kasparov

IBM’s Deep Blue defeated world chess champion Garry Kasparov, demonstrating AI’s ability to surpass human capabilities in well-defined tasks.

IBM Deep Blue

Verified Sources: IBM Sources reviewed:

Key Figures

Geoffrey Hinton

Father of Deep Learning

Hinton is a pioneer in neural networks and deep learning. His contributions to backpropagation, Boltzmann machines, and capsule networks have been groundbreaking. Even during the AI winter, he persisted in neural network research and received the Turing Award in 2018.

Garry Kasparov

World Chess Champion

In 1997, Kasparov was defeated by IBM’s Deep Blue. This event not only demonstrated AI’s capabilities but also sparked philosophical reflections. He later became an active advocate for AI.

James Lighthill

AI Critic

Lighthill’s report had a profound impact on UK AI research, leading to a decade of funding cuts. However, his criticism also prompted AI researchers to be more pragmatic.

Karen Spärck Jones

Pioneer of Information Retrieval

In 1972 Spärck Jones published inverse document frequency in the Journal of Documentation, weighting a term by how rarely it appears across a collection. The idea underpins search ranking and remains inside today’s retrieval systems. She worked at Cambridge through the years when funding for AI was scarce, and argued that computing was too important to be left to men.

Classic Quotes

However, I have to say that, in the AI field, there has been an immense amount of work done, with an immense amount of money spent on it, with an immense amount of hype and excitement. No major results, no grand results have been standing up to the scrutiny of this very searching inquiry, which has been made by a committee of some of the world’s leading scientists.

James Lighthill

Deep Blue was conclusive proof that machines could surpass humans in complex cognitive tasks that we had long assumed were unique to our developed brains.

Garry Kasparov

View Full Timeline →