The Debate: Human Curation or Algorithm?


During a keynote talk at this year’s WSJ.D Live, the technology conference hosted by The Wall Street Journal, Apple CEO Tim Cook defended the marketplace performance of Apple Music, his company’s new(ish) streaming service.

With Apple Music, Cook is betting heavily on human curation, in the form of music experts picking songs for users, as a competitive differentiator. Nor is Apple alone in that endeavor; in a bid to stay relevant in the streaming-music space, Spotify likewise relies on human beings to choose new tunes for others.

Outside of the music realm, Facebook’s new digital-assistant, M, is described as an “artificial intelligence that’s trained and supervised by people,” who can do everything from order gifts to book restaurant reservations. Given Facebook’s size, it’s an open question whether the company will rely more on artificial intelligence, and less on people, as it rolls M out to more customers.

Apple obviously has the money to cycle up its human-curation efforts to gargantuan scope, something that led Alphabet executive chairman Eric Schmidt to term the company elitist. “Today, you’re much better off building a smart system that can learn from the real world—what actual listeners are most likely to like next—and help you predict who and where the next Adele might be,” Schmidt said. “As a bonus, it’s a much less elitist taste-making process.”

But Schmidt can say something like that because Alphabet (the holding company that contains Google) can spend billions of dollars developing sophisticated machine-learning systems that act in increasingly nuanced ways. For smaller app and software developers without those sorts of funds, the question of whether to rely on people or algorithms to provide service is a much more difficult one; people and software are equally expensive, although the latter is cheap to scale up.