Lamenting the incessant 24-hour media speculation over this year’s NBA MVP award, ESPN sports journalist Skip Bayless commented: “I know that’s the age we live in. It’s instantaneous news. We have to be ahead of it. We actually have to make the news before the news gets made.”
They’re no fortunetellers, but media executives are getting closer than ever to offering up-to-the-minute news. In the past few years, the Associated Press, Yahoo and other major news outlets have adopted new policies relegating a portion of their content to high-speed ‘robot journalism’ software. The revolutionary writing programs can produce articles at speeds of up to 2,000 stories per second. When it comes to pace of production and fact mining, writers and computers are worlds apart.
With new changes come new considerations. As they implement alternative technology, publishers need to re-evaluate their ethical writing standards and address the employment implications that automatic content generation will have on writers. The steps used to publish automated content as well as the roles traditional writers will play in the process will both have to be adjusted.
Writing software and traditional journalists find and apply data very differently. Robot journalism technology is undeniably more efficient at compiling and sifting through online information about current events. Quakebot, the geological writing platform used by the LA Times, automatically produces articles about tremors as soon as the U.S. Geological Survey issues notice of an earthquake that meets Quakebot’s minimum magnitude requirement.
Robot journalists can quickly generate articles by using algorithms that collect information from approved sources and plug them into templates. In the case of Quakebot, it’s safe to say that USGS data is typically accurate, but media outlets must remember that it is their ethical obligation to ensure that all sources are similarly trustworthy and their software is collecting the proper information. Ken Schwencke, the creator of Quakebot, admitted that his algorithm occasionally creates articles based on false alerts or technological problems that come up at the USGS. Making sure that editors review articles before they are published will allow the LA Times to provide consistently accurate information.
Robot-generated content is usually statistically rich and short, and it is often used to inform medium-sized or local audiences about financial earnings updates, real estate descriptions, sports recaps and geological news. Given the focus on numbers and facts offered by outside sources, publishers who use robot software should make sure that content is not only correct but also retrieved in ways that do not infringe upon copyright laws. Media outlets, such as the AP, have built a reputation of legitimacy, and that reputation cannot be taken for granted. This is particularly important in early stages of robot journalism, as most readers are still unfamiliar with the workings and use of the writing software.
Digital journalism offers multiplatform content at incomparable speeds, but media representatives should acknowledge that it also puts the employment of some writers at risk. I cannot help but feel insecure comparing portfolios with robots that produce millions of articles each week. As much as I’d like to think the connection writers share with readers – i.e., our ‘it’ factor – is wholly irreplaceable, I don’t. Evolving technology often results in job cuts, an unfortunate reality for employees in myriad fields. Hopefully, managers will do their best to soften the blow by shifting work responsibilities to fact checking and source evaluation.
For years, the writing landscape has been changing, and automated content plays a role in addressing the assignments of modern journalists. Talented writers shine by engaging their audience in analytical pieces, but they are stretched thin by modern demands for rapid content in the form of articles and social media updates. While robot-generated articles aren’t analytically or stylistically imposing, what they lack in panache they make up for in their ability to satisfy the need for speed that underlies contemporary media. Writers who are primarily responsible for smaller-market articles and social media commentary may soon find their job opportunities shrinking.
Competing interests are bound to arise between traditional journalism and media front offices that need to produce quick content without fraying financial resources. It is not unreasonable to think that the introduction of writing software will ultimately threaten the work of writers. The media industry has been facing financial hardships for years. According to the American Society of News Editors, 16,200 full-time newspaper newsroom jobs were lost between 2012 and 2013. Since then, job losses have continued for those working at the Tribune Co. and Time Inc., as reported by the Pew Research Center.
It is possible for publishing managers to alleviate tight budgets by employing increasingly sophisticated automated software. If you think that artificial writing technology with sufficient reader appeal is the stuff of sci-fi, keep an open mind. Robot journalism is not limited to simplistic plug-and-chug work. Although much of it is still focused on writing straightforward overviews and recaps, writing software is becoming adept at implementing natural language.
Article template algorithms can be adjusted to produce tones that appeal to its target audience. They can address readers with an academic or conversational style, and they can mimic human sentiments. When it comes to sports articles, for example, robot journalism software accumulates game data, finds statistical trends, and adjusts content to give off an empathetic or enthusiastic voice.
A sample article used by GameChanger captures the idea with a sports recap opening statement:
Chris Andritsos came up big at the dish and on the bump, leading The Woodlands Highlanders Varsity to a 6-1 win over Austin Bowie on Friday at Mumford, TX. Chris racked up four RBIs on two hits for The Woodlands Highlanders Varsity. He homered in the first inning and doubled in the sixth inning. The Woodlands Highlanders Varsity got the win thanks in large part to Chris' dominant, 12-strikeout performance.
Not exactly a feat of creative genius, but wit and creativity don’t always top the agenda if quantity and speed are a priority – especially when it comes to addressing limited markets. The program did what it needed to. It produced a relatively enthusiastic overview, and a few minutes later, it fired off several more stories at speeds human employees cannot match.
The Big Ten Network and Forbes use Quill, a natural language generation (NLG) platform, to produce similarly straightforward work. The content includes full articles, summaries, tweets and statistics. Whether you look to preferred websites, apps or social media outlets to find the latest information, there’s a good chance you’ve come across its artificially produced content.
More sophisticated use of natural language software has also paved way for robot journalism to enter the world of book writing and poetry. That’s right, writing algorithms can produce long-form content and emotionally charged verse. This isn’t the software’s primary use, but the advanced nature of robot journalism has potential to evolve, leaving less established writers looking for work.
Addressing concerns about job security, AP vice president and managing editor Lou Ferrara told journalists that no jobs were lost as a result of robot software integration: “Automation was never about replacing jobs,” he explained. “It has always been about how we can best use the resources we have in a rapidly changing landscape and how we harness technology to run the best journalism company in the world.’’
Some professional journalists share Lou Ferrera’s optimism about the adoption of automated writing software. I’m tempted to call their bluff, but perhaps their conviction about job opportunities is heartfelt. There are optimists in every field.
For New York magazine writer Kevin Roose, the introduction of robot journalism is a welcome one. Roose is relieved that journalism software can be used to complete assignments most humans hate. He recalls dreading being pulled from his work to complete recaps of corporate earnings: “The stories were inevitably excruciatingly dull, and frankly, a robot probably could have done better with them,” he writes. Schwencke also finds the software to be harmless: “It’s supplemental,” he told Slate magazine. "The way I see it is, it doesn't eliminate anybody's job as much as it makes everybody's job more interesting."
Essentially, established writers view robot journalists as overqualified interns who don’t complain about the workload or pout about their edits. And if you’re already a seasoned writer, it’s not unreasonable to think that way. Experienced, highly talented journalists will always be needed for their analytical skills and personable writing.
As of now, writing software can’t pull off uniquely cerebral content or conduct interviews that give writers material for poignant profiles and cultural pieces. Robot journalism is nowhere near acing interactive work, but the digital landscape is filled with little-known writers whose work depends on the more monotonous part of writing. Media outlets should truthfully address the future of employees whose work is increasingly produced by algorithms. Optimally, they can utilize current writers to ensure that automated content maintains their standards of accuracy.
What is the role of behind-the-scenes writers in an increasingly automated media landscape, and how will publishers maintain their ethical responsibilities as content producers and employers? According to Automated Insights CEO Robbie Allen, the company’s writing platform is already the largest producer of narrative content from big data in the world. If executives are looking for numbers, then writers looking to contribute to the workforce face an uphill battle. But if employers continue to set a high bar for management and content, there may be a light at the end of the tunnel.