Robot call centers. Robot cars. Robot nannies. Robot dogs. Robots making robots. The list goes on…
Businesses are using robots to increase efficiency and improve their bottom line. Recent research has found that businesses leveraging automation tools for marketing tasks have a 53 percent higher conversion rate than peers who stick to purely human labor.
Robots are now on the verge of disrupting yet another industry: journalism.
You may not realize it, but you’ve already read content produced by robots. Believe it or not, these robo-journos have become prolific in a few fields, including finance and sports.
Is it that hard to imagine a bot that spits out cookie-cutter data, like how much a particular stock rose or fell on a given day? Or even an automated report on your favorite sports team’s latest win? Many stories have a determined structure, with interchangeable details. Why not assign a robot to the beat?
After all, someone—or something—had to replace all the journalists and reporters who lost their jobs during the recession.
Thanks to algorithms created by companies like Narrative Science and Automated Insights, this variety of automatic content can be generated without much effort by expensive human employees. While the results might not be Pulitzer material, what’s written is perfectly readable and (almost) wholly sensible.
Though the idea that powers this kind of technology has been around for nearly 40 years, it hasn’t been utilized on a widespread basis until recently. Even the technology’s harshest critics are at least somewhat impressed by its results.
So how does this automated journalism technology work, anyway?
For starters, these robo-journos need to be fed a lot of data, which serves as the foundation of their intelligence. That data needs to be very accurate, and companies need to ensure as little human error, if any, is involved in the process.
Next, the robo-journo needs to be able to distinguish the important facts and newsworthy nuggets hiding in every data set it digests. The technology is currently being used in finance and sports for simple reasons: Both fields involve scores of recurring and patterned data. A baseball team can either win or lose, assuming the game doesn’t get suspended or rained out. A stock will either increase in value, decrease in value, or stay the same price.
It’s not hard to see how routine stories are easy for robots to write. But a present-day robot would have a hard time suitably recapping the carnage that resulted from an all-out bench-clearing brawl, with fans jumping onto the field and ballplayers attacking each other for 20 minutes. For the time being, only a human can accurately report on such irrational behavior!
The Impending Fall of the Robot Journalist
Robot journalists can serve a purpose. But just like humans, robots make errors, too.
Case in point? In 2014, the Wikipedia Live Monitor—a bot that scans Wikipedia edits across languages to synthesize the breaking news of the day—sent out a tweet saying that an article about Quinton Ross, an NBA player, might very well qualify for breaking news.
It was true that a man by the name Quinton Ross had died. But it wasn’t the basketball player, as the New York Post erroneously reported earlier that day. As humans began retracting and correcting their stories, the Twitter bot remained silent, unaware that it had spit out misinformation.
The concept of the robot journalist boils down to this: It is a high-tech algorithm that can scour the Web to find what’s being talked about, digest that content, and spit out a legible analysis.
Do you see what’s missing from this equation? The ability to ask questions, make abstract connections and, of course, inject a little bit of personality into the words.
Does anyone really want to read robot-generated content that might also be factually incorrect? And if it is incorrect, who will fix it? Who will be held responsible?
While we’re bound to see more content generated by robot journalists in the future—especially as the technology becomes even more refined—it’s more than likely that these articles will serve to supplement real, human-written stories—not supplant them.