Skimming along the waves of the world wide web a wee while back, I chanced across a rather disturbing idea, for anyone who’s ever watched Terminator: robot news. By this I don’t mean a live blog of R2-D2′s address to the American people – I’m talking about using StatsMonkey; an automated bit of software made by Intelligent Information Laboratory to create journalism from baseball statistics.
For me, this hits uncomfortably close to home. I’ve been working in journalism since I was seventeen, beavering away on subs at the Financial Times. It was hard graft there, and even more so when you were on the phone to Dow Jones, hoping they took you seriously with a high-pitched, quivering voice. But the experience was solid, and it helped me further my interest in bringing the news to the people.
However, if I was to suggest that financial journalism, the most statistic-laden, number-driven form of documenting current events, was to be automated, they’d have laughed me out of the office. “You can’t document the world of finance with a bit of code,” they’d claim. “You need real people, with real experience and talent.”
But how talented are we, really? I’ve seen robots play the piano, and though many claim it sounds lacklustre, without knowing better they’d think it was Mozart himself. There are many programs that can actually converse with you, they respond to what you’ve said with pre-programmed syntax strings and match your concepts to their comments.
So what makes these cocky programmers think they can master the craft of the written word with numbers and bits of fiddly code? Well, being a journalist, I went and had a dig around, as getting hold of these mad journo-scientists seems to be harder than curing the common cold (not that I’d know where to start, bar recommending ice cream, movies and a few days of rest).
The first piece I saw was Wired. Honestly? I thought they’d be massively against it. Surely using sports data to compile a story is no more difficult than doing the same with new technology? Cross-referencing previous scores/specifications, talking about other teams/brands, and comparing them to the matches/devices of legend.
But I was wrong – they embraced the idea and openly stated it would be cheaper. There was more talk of software writing a Pulitzer-Prize-winning book, but let’s stay out of science fiction predictions for now. The realistic side to this development is this – journalists, and especially news journalists, are simply regurgitating facts in a certain format.
This is highlighted by National Public Radio’s interview with one Kristian Hammond, co-director of Intelligent Information Laboratory. When asked about documenting Little League games, he responded:
“We could literally write a game story for every single Little League game that is played in this country. That means every kid, every dad, every family, every grandma would see the story of what their kid is doing.”
The potential of this technology is huge. The ability to document every sports game, every bit of company progress – essentially, anything involving data, is mind-blowing. I’ve had instant-message chats with a computer, and it works to a point. But news? News is formulaic.
You can call me on it all you want, I’ve done my time in journalism – news with nothing but facts is news. News with opinion is a column. News with StatsMonkey, however, might be the reason none of us have a job in ten years. Medical degrees, anyone?