In an article published last weekend on Mashable, Sarah Kessler asked the question, “Can Robots Run the News?” It’s an important question not just for journalists, but for anyone who creates or curates content on the Web.
The examples Kessler cites span the range of content creation, from automatically generated sports news to the use of algorithms to identify news topics. There’s obvious value to automated content creation, and as Jeff Jarvis has declared, “Data is (are) journalism.” But we should be careful not to confuse computed content with communication.
Computed content is a set of data; communication is the expression of an attitude toward, or perspective on, those data. Without a point of view, content is just an audience speaking to itself.
Using Web analytics from a test period to automatically choose between two headlines, as we’re told the Huffington Post does for its stories, can make sense—if both versions are true to the content. If you balance crowd-sourced feedback with the content creator’s point of view, you’ll have a productive conversation. But if the crowd takes precedence, it may simply replace content’s individual vitality with the bland mean.
Take, for instance, the English title for Stieg Larsson’s novel The Girl with the Dragon Tattoo. It may not have been crowd sourced, but it certainly plays to a corporate idea of the crowd. Is it really better than the literally translated original title, Men Who Hate Women? (That’s a rhetorical question. The original title nails the book’s central concern; the English version just wraps it in a pulp-fiction cover.)
Even in content marketing, where knowing what people want is critical to the content provider’s success, a one-sided conversation dominated by the audience won’t fly. For a conversation to work, there must be differences between the participants. The power of new media is the way it enables the audience to challenge the creator. That doesn’t mean, though, that the creator should stop challenging the audience.
This balance seems to be what Yahoo VP of Media Jimmy Pitaro is after in the company’s news blog, The Upshot. In her interview with him last week on All Things D, Kara Swisher noted that while some see computational journalism as a “‘democratizing’ of the news, others are more concerned about relying on algorithms to determine the best coverage and the implications for a society guided by its own searches.”
But as Pitaro noted in his video interview, “data and audience insights” constitute just one component of the content. In addition, Yahoo uses the “old-school” methods of “manually identifying topics” through its team of editors and writers.
Similarly, as Kessler mentioned in Mashable and as Claire Cain Miller explored at greater length in yesterday’s New York Times, the tech-news site Techmeme uses both algorithms and editors to produce its content. Why? Because “humans do things software cannot, like grouping subtly related stories, taking into account sarcasm or skepticism, or posting important stories that just broke.”
If readers didn’t care about such things, algorithms alone might be enough. But they do care. The same audience whose searches drive the algorithms also want the human touch in their content. Until computers can pass the Turing Test, it isn’t likely that they will replace people in content creation.