rss feed

writing in the age of robojournalism

As the story goes, an earthquake struck Los Angeles in the wee hours of February 1, 2014. Within eight minutes, a content generation algorithm posted the story to the L.A. Times, including a map pinpointing the epicenter. This wasn’t the first time a bot was used to write a story (some say this happened in the early 90s) but, for many people, it was the first time “robojournalism” became applicable to real life.

Unlike Ken Schwencke, the programmer/journalist who built the Quakebot software, most professionals who write for a living probably don’t have the time or desire to build software that automates their own tried-and-true process for researching, writing, editing, or fact checking. But with so-called experts predicting that content generation programs will replace human writers, should writers fear AI, data analytics, and machine learning? Or can content generation programs actually benefit human writers? So, I set out to unravel some common myths.

    • MYTH #1: “Content generation programs are more efficient than humans” – In some cases, absolutely. There are tools that mine social media for breaking news. Natural language processing services that perform audio transcription with relative accuracy. And, like Quakebot, story-assembling software solutions that reduce time-to-publishing for news rooms, marketing teams, and others.

      The thing is, efficiency is just one aspect of successfully writing on deadline. Accuracy is just as important, and content generation platforms require journalists, marketers, and other skilled writers to provide the parameters and templates to train them. After all, “garbage in, garbage out.” Then there’s relevance—whether a story matters to a given audience. While these platforms can estimate which topics matter to which demographics and even write data point-centric stories, they lack the original reporting skills and nuanced experience needed to write compelling stories that captivate and influence readers.

      • MYTH #2: “Audiences can’t discern between human- and bot-written content” – Well, can you? The Washington Post published an article last summer to see if readers could tell the difference between human- or bot-created content. If you haven’t, I recommend you take the quiz. Full transparency: I only got 4 of the 7 written samples correct. What’s more, many businesses are using content generation programs to crank out short, data-centric, surprisingly-tailored content—from the Associated Press to Yahoo Fantasy Sports to L’Oréal.

        The results become less convincing once you try to convey complex topics, dependencies, or emotions that can’t simply be learned through structured data points. Like Hollywood asserts, bots have a hard time understanding human emotions beyond predictable clues. According to a study by predictive intelligence service Motista, emotional connection and customer purchase spend have a direct correlation. But, if bots can’t write with human-like emotion, human intervention is crucial. Are you willing to entrust writer bots with your entire customer journey? I’m not.

        • MYTH #3: “Robots are going to replace human researchers, writers and editors” – There are dozens, if not hundreds, of content generation offerings available today, from text-to-speech and sentiment analysis point solutions available through AWS, Azure, or Google; to trainable, in-depth content generation programs like Wordsmith, Yseop, and Heliograf that can produce thousands of articles or reports a week. But these programs can’t replace some of the key cognitive and interpersonal functions human writers perform, like interviewing and understanding subject matter experts, attendees, or witnesses, not to mention employing nuance to structure a narrative that will resonate with a specific audience. So, no, human researchers, writers, and editors’ jobs aren’t necessarily at risk.

          If anything, these technologies are an opportunity wrapped in a question: How can I do my job better and bring better value to my organization? Like what many at The Washington Post are already doing, these technologies can help writers perform the basic functions of their jobs faster and easier, so they can focus on what you probably like best: simplifying complex ideas beautifully and influencing audiences to think and act differently.

Net net: Writing isn’t a commodity, but a writer’s role isn’t static either. It’s up to savvy writers—and the businesses that work with them—to continuously elevate their craft and trade…even if it means befriending a few bots along the way.

rss feed