Is A.I. “Intelligent” Enough? The Case for a Human (and artificial) Touch

Illustration by Tom Cheney

In our new digital age, every time a story breaks, you can be sure the entire world will react. Even if the only thing you do is log into Facebook or Twitter, you’re guaranteed to be greeted with a stream of stories on the topics du jour, from politics to technology to health. If we’re not careful, it’s all too easy to fall victim to the echo chamber and encounter only the same select few facts, interpretations and perspectives over and over again.

It’s a tougher problem to overcome than you might think. Even though there are literally hundreds of thousands of sources of original content out there in the world today, it’s always the same few dozen voices that rise to the top of Google News. There is an intense pain that all of us who are information-seekers feel: we have a burning desire for high-quality, diverse, and relevant information, but only a limited amount of time and resources to put into acquiring it.

Maybe this is a good problem for humans to solve? We’re outstanding at discriminating between high-and-low quality articles and publications; we can recognize nuance and thoughtfulness and separate out shills and demagoguery in mere seconds. The expertise we’ve acquired in our fields allows us to recognize a brilliant piece among a sea of mediocrity in ways unmatched by any automaton. Yet we’re also extraordinarily limited by how long it takes us to read and process information. Even if we sift through a thousand articles a day from the sources on our radar, there’s still a great chance that we’re missing out on some stellar content from a source we haven’t encountered before, or one we don’t associate with the type of content we’re seeking out.

So, then, maybe we should outsource this problem to a machine? After all, machines can quickly aggregate, sort and filter huge quantities of information, and select or flag articles based on whatever criteria you program into them. The very parts of the problem that are most difficult for humans are a snap for machines. And yet, despite everyone’s best efforts, machines on their own are still easily duped by SEO tactics and are unequivocally lousy at distinguishing articles based on quality.

Artificial intelligence is an incredible development, but it’s still no replacement for a human’s expertise. The way to optimally solve this problem is to first let the machines do the part they’re good at: aggregate and scan the articles as they’re created, sort and filter them according to an intelligent algorithm, and return only the best content to select from. But then, get a human, preferably one with the right expertise, to accomplish the tasks that they’re good at: to pick out the nuance, to select articles for quality and relevance, and to choose a diversity of perspectives to present to the final audience.

Don’t let the sheer amount of content overwhelm you, but don’t fall into the trap of forcing a machine to perform a task it’s not suited for. It’s only through the fusion of man and machine that we can ensure that the right content reaches the right audience at the right time.

-Ethan