A week or so back the New York Times published one of those articles that must have enlivened the lunchtime conversation among the paper’s staff reporters. Steve Lohr wrote about a company called Narrative Science, which claims to have developed computer software that can write news articles. According to Lohr’s piece, the software ingests sports or business facts and cranks out articles that sound as if they’ve been written by an actual human being, all for the absurd price of 500 words for less than $10, a rate you can’t match in Bangalore. In Lohr’s words:
For years, programmers have experimented with software that wrote such articles, typically for sports events, but these efforts had a formulaic, fill-in-the-blank style. They read as if a machine wrote them….But… the articles produced by Narrative Science are different.
Lohr goes on to explain how it works:
The Narrative Science software can make inferences based on the historical data it collects and the sequence and outcomes of past games. To generate story “angles,” explains Mr. Hammond of Narrative Science, the software learns concepts for articles like “individual effort,” “team effort,” “come from behind,” “back and forth,” “season high,” “player’s streak” and “rankings for team.” Then the software decides what element is most important for that game, and it becomes the lead of the article, he said. The data also determines vocabulary selection. A lopsided score may well be termed a “rout” rather than a “win.”
The clincher comes when Lohr speculates on “a media maven’s prediction that a computer program might win a Pulitzer Prize in journalism in 20 years” and then quotes Kris Hammond, one of the Narrative Science founders.
“In five years,” he says, “a computer program will win a Pulitzer Prize — and I’ll be damned if it’s not our technology.”
Well, I’ve taken the trouble to read a Narrative Science robot’s account of the University of Wisconsin’s opening win of the football season over UNLV and I’ll be damned if a bot is ready to contend for the prize in 20 years. In fact, if this dude (can bots that write be called dudes?) wins a Pulitzer in 20 years, I’ll eat my mousepad.
Lohr’s article offers minimal quotation from the sportswriting bot – I think anything more extensive in the way of bot quotes would queer the little game as provocateur that he’s playing – but I looked up the bot’s story of the game and read it next to one written by a real sportswriter named Tom Mulhern, whose story appeared in the Wisconsin State Journal and Capitol Times and on the newspaper’s site, Madison.com.
You be the judge. Here are the first seven paragraphs of the bot in action:
Wisconsin jumped out to an early lead and never looked back in a 51-17 win over UNLV on Thursday at Camp Randall Stadium.
The Badgers scored 20 points in the first quarter on a Russell Wilson touchdown pass, a Montee Ball touchdown run and a James White touchdown run.
Wisconsin’s offense dominated the Rebels’ defense. The Badgers racked up 499 total yards in the game including 258 yards passing and 251 yards on the ground.
Ball ran for 63 yards and three touchdowns for the Badgers. He also caught two passes for 67 yards and a touchdown.
Wilson completed 10-of-13 passes for 255 for Wisconsin. He threw two touchdowns and no interceptions.
Caleb Herring threw for 146 yards on 18-of-27 passing. Herring tossed two touchdowns and no interceptions.
UNLV had 292 total yards. In addition to Herring’s efforts through the air, the running game also contributed 146 yards for the Rebels.
Sounds a little like the folks who wrote the Dick and Jane reader series, authors of the immortal line: “See Spot run.” Now here are the first seven paragraphs of Mulhern’s story:
Despite all of the attention focused on transfer quarterback Russell Wilson, the running game is still the engine that drives the University of Wisconsin football team’s offense.
The Badgers reminded a national television audience of that in another dominating offensive performance in their opener on Thursday night, thrashing UNLV 51-17 in front of a crowd of 77,085 at Camp Randall Stadium.
Many fans undoubtedly left talking about the impressive debut by Wilson, who completed 10 of 13 passes for 255 yards and two touchcowns. He also ran twice for 62 yards and another touchdown.
Yet, it was set up by a strong effort from the running back tandem of junior Montee Ball and sophomore James White. Along with a top performance by the offensive line, it made for an easy night for Wilson, who was hardly touched by a UNLV defender.
“We feed off of that,” UW coach Bret Bielema said of the running game. “(Wilson) really saw how that can open up things in the run and play-action game.”
The Badgers viewed the nationally televised game, the unofficial start to the college football season, as a chance to make a statement to the rest of the nation.
That statement is, this offense could have a chance to be even better than the one that set the school record a year ago, averaging 41.5 points per game.
I’m going to presume that as a reader of 317am you are savvy enough that I don’t have to belabor the point about the minimalistic crudity of the bot’s prose style and the comparative virtues of Mulhern’s more analytical account. I’m not quite ready, however, to write off Narrative Science as a crew of mountebanks.
I know developments in information technology seem to be moving at an ever-increasing rate of speed. I can imagine that as these programs become more sophisticated and more actual writers – probably dead ones whose copyrights have expired – get programmed into the fold, some day you will be able to order up your sports news as crafted by your favorite writer.
Want that Wisconsin-UNLV game a la Montaigne or Jane Austen? No problem. A Dr. Johnson version as he might have dictated it to Boswell or in his comprehensive Dictionary format? Both will be available. The Deluxe William Faulkner Edition of the Lost Cause Game Summary? If you’re willing to pay a little more, of course.