PCMag editors select and review products independently. If you buy through affiliate links, we may earn commissions, which help support our testing.

Will Computer-Generated Articles Replace Human Journalists?

 & Leslie Horn Reporter

Our team tests, rates, and reviews more than 1,500 products each year to help you make better buying decisions and get more from technology.

Our Expert
LOOK INSIDE PC LABS HOW WE TEST
65 EXPERTS
43 YEARS
41,500+ REVIEWS

A group of researchers in Illinois have come up with software that automatically generates news stories. But will it replace human journalists?

The New York Times reports that the program, which turns data like sports stats, financial reports, and housing information into stories, is the result of more than 10 years of research from Narrative Science, a startup in Evanston, Ill. It's headed up by Kris Hammond and Larry Birnbaum, directors of Northwestern University’s Intelligent Information Laboratory. While the articles are created using smart software, they don’t read like they were penned by robots.

“The quality of the narrative produced was quite good,” University of Washington computer scientist Oren Etzioni said. Etzioni told the Times its evidence of the computing trend of “the increasing sophistication in automatic language understanding and, now language generation.”

According to its Web site, “Narrative Science transforms data into high-quality editorial content.” It is able to produce “news stories, industry reports, headlines and more – at scale without authoring or editing.”

In the past, experiments with this computer-generated journalism have yielded work that sounded mechanical. But Narrative Science’s stories sound like they were written by actual people.

Founded last year, the company told the Times it already has 20 clients, including the Big Ten Network, which started using Narrative Science in the spring of 2010 to write short summaries of baseball and softball games. The Times said the technology generated stories that were posted on the Big Ten Network Web site within a minute of the end of every game.

The Big Ten Network worked with Narrative Science to help the software learn different sports concepts, generate angles, and master jargon. The software culls through data to determine the most important information, and leads with that element. The network said articles also improved over time, as Narrative Science used historical data to understand the sequence of games and refine language choice.

“Composition is the key concept,” Hammond said. “This is not just taking data and spilling it over into text.”

But journalists need not worry about losing their jobs to this software. The Times said the software has mostly been used by news outlets working on a limited budget that are looking to boost their coverage or papers looking for summaries of youth sports or local companies’ financial reports. 

About Our Expert

Leslie Horn

Leslie Horn

Reporter

Leslie Horn joined the PCMag team as a news reporter in the fall of 2010. She covered a wide range of topics, from digital media to the latest Apple rumor. After graduating with a degree in Magazine Journalism from the University of Missouri, she wrote for Out & About, a travel guide in coastal Maine. One of her favorite reporting experiences was covering the 2008 Olympics from Beijing. She travels every chance she gets; a favorite trip was backpacking along the coast of Brazil. Though she was born and raised in Dallas, Texas, Leslie embraces life as a New Yorker.

Read full bio