You can’t build a useful skyscraper without elevators. In much the same way, the AI breakthroughs we see today wouldn’t exist without the foundational technologies developed during a quieter chapter in AI’s history.
From the late 1970s to the early 2000s, AI experienced a lull in mainstream attention. However, behind the scenes, essential components were being developed, refined, and widely adopted. These technologies would later make modern AI possible.
Building the Foundation
This era saw the rise of microcomputers, laptops, mobile phones, and early tablets like the PalmPilot. Hardware became smaller, faster, and more accessible. At the same time, sensors capable of detecting heat, digital cameras for both still and moving images, and digital audio tools for recording and editing began to spread.
Key Capabilities Emerged, Including:
- Natural Language Processing (NLP): essential for search, grammar, syntax and translation
- Optical Character Recognition (OCR): allowing machines to read printed text
- Optical Mark Recognition (OMR): enabling systems to detect marks or shapes
- Speech recognition: converting spoken words into text
- Image and face recognition: distinguishing between objects, people, and scenes
Alongside this, infrastructure developed rapidly. Databases began storing and indexing images, videos, and text. Natural language processing improved. The internet connected everything, giving rise to digital communication through email and web content. This explosion of text and data formed a searchable, usable knowledge base that would later become fuel for machine learning models.
Quiet Breakthroughs with Big Impact
In the late 1980s, Germany produced the first driverless car. A decade later, Akash Deshpande at UC Berkeley advanced the concept by piloting five autonomous vehicles through the streets of Los Angeles.
Meanwhile, IBM developed Deep Blue, a chess-playing computer that initially won one of six matches against world champion Garry Kasparov. In a 1997 rematch, it won outright by analyzing 200 million moves per second (far beyond human capacity).
By the end of the 1990s, the groundwork was nearly complete. Google was founded in 1998, bringing new power to search and information access. Around the same time, Dr. Cynthia Breazeal at MIT created Kismet, a robot capable of expressing and recognizing human emotions.
Continue watching and reading our series on the history of AI: