Douglas Lenat, an artificial intelligence researcher formerly of Wyncote, will always be remembered for his pioneering work in AI, writes Cade Metz for The New York Times.

Lenat, 72, passed on August 31 in Austin, Texas from bile duct cancer.

During his time as a professor of computer science at Stanford University in California, he discovered his passion for AI technology. In the late 1970’s he developed an AI system called Eurisko which he designed to automate the discovery of new scientific concepts and methods.

Then in 1981, he used that system to analyze the rules of a role-playing game called Traveller Trillion Credit Squadron to find new ways to win the game.

In the game, players design and deploy warships. Eventually, he identified winning strategies that allowed him to win a Traveller tournament in San Mateo, California.

His experiences with Eurisko made him realize that machines could not be truly human without common sense, so he began a new project called Cyc to build common sense by hand, one line of code after another.

To do this, he needed to define the fundamental laws that govern how the world works like “you can’t be in two places at the same time.” He knew it was a massive undertaking that could take decades, but he believed strongly in a rules-based approach to AI.

In more recent years, AI researchers have taken a different approach, building systems that can learn skills by analyzing large amounts of digital data. ChatGPT is one example of that strategy.

AI experts have mixed views on chatbots. Some believe that they will eventually be able to duplicate human reasoning, while others are more skeptical.

“These chatbots think that when you hammer a nail into the wall, it should be vertical,” said Ken Forbus, a professor and AI researcher at Northwestern University in Illinois.

“They can be very useful. But they don’t understand the world.”

Read more about Douglas Lenat and his life work influencing the evolution of AI in The New York Times.

_______________________

More on the evolution of AI.