Automation's next frontier: journalism | ZDNet
Many newspapers already have applications that gather sports scores and feeds them into legible, machine-generated articles on the latest games. Now, a startup is ready to take this technology to the next step -- financial reporting and beyond.
MarketWatch's Tom Bemis reports on a startup that markets a solution that converts streams of data into readable articles:
"By feeding a set of numbers through its systems, the two-year-old Narrative Science can generate stories in English that can explain the data from charts, graphs and spreadsheets to their customers... One straightforward application is 'looking at point-of-sale data for a franchise organization ... or any organization that has corporate headquarters and individual outlets,' [says Kristian Hammond, CTO and co-founder of Narrative Science.]
Narrative Science's artificial intelligence platform, Quill, is "designed to understands the stories it writes, independent of the structure and language in which it expresses them. As a result, it can use the same set of ideas to generate stories in multiple formats, from long-form stories to PDF’s to business reports to Tweets. Quill also understands the relative importance of the elements within a single story and across multiple stories."
The value proposition isn't replacing human writers, but actually enabling the writing of stories targeted at extremely small audiences -- where paying a writer wouldn't make economic sense anyway. “The place where it’s most powerful is where we have an audience of one,” Hammond is quoted as saying. “Those stories tend not to be written by people, but they can be written by a machine.”
Interestingly, and perhaps tellingly, Hammond also is a professor of journalism and computer science at Northwestern University. Narrative Science started life as a joint research project at Northwestern University Schools of Engineering and Journalism, and its first automatically generated story described a Northwestern Wildcats baseball game.
(Photo: Joe McKendrick.)
This post was originally published on Smartplanet.com