This is Andrew Ng’s talk at the Stanford MSx Future Forum, delivered on 25 January 2017 and posted on the official Stanford Graduate School of Business YouTube channel. At the time Ng was chief scientist at Baidu, a co-founder of Coursera, and an adjunct professor at Stanford, and the talk has become one of the most-cited popular framings of modern AI.
Ng’s central analogy is that AI is the new electricity: just as electricity transformed transportation, manufacturing, agriculture, and health care roughly a century ago, he argues he has a hard time naming an industry that AI will not reshape over the following years. He grounds this in supervised learning - mapping inputs to outputs - and explains why more data and bigger models drove the gains of the deep-learning era.
He is also deliberately measured about the limits. He stresses that AI remains narrow and extremely limited relative to human intelligence, that the main bottlenecks are data and talent rather than algorithms alone, and that fears of imminent human-level machines are misplaced. This makes the talk a useful beginner-level primary source: a firsthand, business-minded explanation of both the promise and the boundaries of the technology.