Pearl Writes Book on How AI Should Take New Direction

Judea Pearl says that for AI to become truly intelligent, the machines must be able to question “why.”

Pearl, a pioneer in AI, recently wrote “The Book of Why: The New Science of Cause and Effect,” which argues the AI has been handicapped by an incomplete understanding of intelligence. Most breakthroughs today are in machine learning and neural networks, which Pearl doesn’t find impressive. Pearl won the Turing Award in 2011.

At 81, Pearl is a critic of AI as it is devised now, saying that programmers need to replace reasoning by association with causal reasoning. Instead of correlating fever with malaria, machines need to know that malaria causes fever. With the right framework for questions, computers will be taught to think probabilistically.

In a lengthy interview in Quanta magazine, Pearl discussed how science has “abandoned” cause and effect in solving problems.