July 12, 2012

Computer Algorithm Reporter Could Produce the News

Peter Suciu for - Your Universe Online

Various themes of the infinite monkey theorem states that if one takes a monkey and a typewriter — or even a thousand monkeys and a thousand typewriters — and gives an infinite amount of time that in turn it could result in the complete works of William Shakespeare. While this has been all but surely debunked — at least until TV´s Mythbusters decide to devote the next thousand years to it — what is certain is that no amount of monkeys and keyboards could ever replace a human news reporter.

On the other hand a computer algorithm could actually do what even a million monkeys working on a million keyboards couldn´t do, namely write news copy. The result could be a new reporter who could turn blazing fast copy that is clean, free of typos and generally grammatically correct.

In fact algorithms are being used to produce articles already for newspapers and websites. One such example is Narrative Science.

In fairness the “reporter” isn´t really doing reporting, but rather creating encapsulated versions of longer news stories. The computer programs thus can´t determine what is actually newsworthy in a given story, nor can it set a tone, but what it can do is take the raw data and create a quick, yet formulaic, news post.

Steven Levy described the technology in a recent Wired magazine column, “The computer-written product could be a pennant-waving second-half update of a Big Ten basketball contest, a sober preview of a corporate earnings statement, or a blithe summary of the presidential horse race drawn from Twitter posts.”

In other words the computer algorithm can write the sort of copy that reporters often “knock out” in a few minutes, and which often times carries a staff report byline. By having the computer write this sort of copy it essentially frees up reporters to do actual“¦ well, reporting.

But is it taking jobs away from journalists at a time when journalism jobs aren´t all that stable? Narrative Science CTO and cofounder Kristian Hammond told Levy that this isn´t the intention.

Instead, Hammond suggested that it could expand news coverage while producing “ultracheap” but “totally readable accounts of events, trends, and developments that no journalist is currently covering.”

This also isn´t the first time that computers have “assisted” in reporting. Computer systems have already been used to sift through data to provide details to human reporters.

“I don't have a philosophical objection to that kind of writing being outsourced to a computer, if the reporter who would have been writing it could use the time for something more interesting,” Stephen Doig, a journalism professor at Arizona State University told wire service AFP.

Narrative Science, which was founded in 2010, now has numerous mainstream clients including Forbes. The company also has some corporate clients, where its technology is used to make spreadsheets and internal reporters more readable. But at the end of the day the goal is really about turning copy quickly, so that again reporters, fact checkers and editors can be freed up to other reporting, editing and checking.

But in the end this approach to news writing is still about data not really reporting.

“We're about two-thirds engineering and one-third journalism,” Hammond told AFP. “We knew there were places in traditional journalism where raw data was used as the driver for telling stories, and we wanted to take that model and turn it into something a machine can do.”