« »

Tuesday, December 30, 2014

How artificial intelligence writes news

artificial intelligence and news
Automated news is made possible with algorithms and data transformation

The story of humans versus robot-computer has reemerged from the realm of information technology with editorial practices firmly fixed in its focus. Some of the news already being delivered to consumers is auto-generated using technology designed by an innovative company named Narrative Science. According to Kris Hammond, Chief Technology Officer of the business, news is created when the algorithms “...take data and transform it in to stories and insight...completely automatically, with no human intervention.”


The method used to create otherwise human generated communications involves appropriately compiling data, then breaking that information down in to its component parts for re-configuration as a news story. This configuration process is accomplished by using a technique called “angular characterization.” Moreover, this methodology matches pre-programmed “angles” such as a home team victory narrative with relevant data. In this example, an identifiable win by one team over another would be expressed via the narrative template by pulling from a suitable phrase database.


News written by algorithms produced by Narrative Science is literate, accurate and even analytical. Furthermore, according to the business' CIO, the characterization of extracted data is outputted as headlines, narratives, earnings previews, Tweets and reports. Furthermore, the Securities Technology Monitor states a software developed by Narrative Science called “Quill” frames information in terms of “best practices,” “specialized version templates,” and “agnostic data analysis.” In other words, it customizes the content for audience beliefs, objectives and preferred styles of presentational tone.


Andy Rooney, the late 60 Minutes reporter once said the following per Brainy Quote: “Computers may save time but they sure waste a lot of paper. About 98 percent of everything printed out by a computer is garbage that no one ever reads.” Whether or not those print outs were derived from information typed in by a human or a computer is up for debate. In other words, regardless of which news is better, over time the advantages of a high-capacity news generating program outweighs the pension, salary and overhead costs required by numerous news reporters and “level one” analysts.


At a rate of thousands of stories per week, intuitive fore-thought still has an edge over narrative engineering per Securities Technology Monitor's interview with a capital management firm executive. Journalists, writers and narrators will be challenged by technology that makes a large portion of basic news delivery efficient. This means, more streamlined corporate operations with lower costs at the expense of the traditional news industry. Any implications on society, culture and human functionality are possibly grouped with the word ontology, which is banned from being used by the CIO of Narrative Science.

Image: Geralt/Pixabay; US-PD