In the 1970s and 1980s, artificial intelligence (AI) made major strides, marked by the rise of expert systems and periods of stagnation known as AI winters. It was an era of breakthroughs and setbacks that shaped the course of AI research and applications.
Rise of Expert Systems
Expert systems are AI programs that mimic the knowledge and reasoning of human specialists in specific domains. They use collections of rules to solve complex problems or make diagnoses.
MYCIN (1972): Medical Diagnostics
One of the first successful expert systems was MYCIN, developed in 1972 at Stanford University. MYCIN was built to diagnose bacterial infections such as bacteremia and meningitis and recommend appropriate antibiotics, with dosages adjusted to a patient’s body weight. The system used roughly 600 rules and asked users a series of questions to reach a diagnosis.
Studies showed MYCIN’s recommendations were acceptable in 69% of cases—on par with human experts. Even so, MYCIN never made it into clinical practice, largely due to legal and ethical concerns and the technological limits of the time.
Stanford Cart (1979): First Self-Driving Vehicle
In 1979, Hans Moravec at Stanford University developed the Stanford Cart, one of the earliest autonomous vehicles. It navigated obstacle-filled spaces on its own, using a video camera and a computer to analyze its surroundings and make decisions.
Movement was slow, and processing between frames took minutes, but the Stanford Cart proved autonomous navigation was possible—laying the groundwork for future self-driving technologies.
AI Winters: Periods of Pullback
Despite early wins, AI also hit periods of disillusionment and funding cuts, known as AI winters.
First AI Winter (1974–1980)
By the 1970s, AI’s limits were becoming clear. The 1973 Lighthill Report criticized the lack of tangible progress, leading to reduced government funding in the UK. This era of skepticism and cutbacks is known as the first AI winter.
Second AI Winter (1987–1993)
After a brief revival in the early 1980s fueled by expert systems, disappointment set in again. Expert systems proved costly to maintain and too inflexible for changing conditions. The result: a second AI winter, with declining investment and interest.
Key Developments
Even amid headwinds, the era delivered important technological breakthroughs.
Fifth Generation Computer Systems (1982)
In 1982, Japan launched the Fifth Generation Computer Systems project (FGCS), an ambitious push to build advanced computing technologies using AI. The goal was to create machines that could reason and learn like humans, with applications from natural language processing to image recognition.
While it ultimately fell short of its lofty promises, FGCS spurred global research and investment in AI.
Rediscovery of Backpropagation (1986)
In 1986, the backpropagation algorithm was rediscovered, reigniting interest in neural networks. Backprop enabled more efficient training of multilayer neural nets, paving the way for modern deep learning. This shift proved crucial for later advances in speech recognition, image processing, and other AI applications.
Conclusion
The 1970s and 1980s were a mix of momentum and setbacks for AI. Expert systems like MYCIN and breakthroughs like the Stanford Cart showcased AI’s promise, while the AI winters underscored the complexity of human intelligence—and the challenge of replicating it. This period laid the groundwork for AI’s later evolution and remains a defining chapter in the history of technology and science.
Loading