A company that has already automated much of sports reporting is set to take the final step towards an artificial journalist. Statsheet collects sports statistics and puts them into easily sorted, manipulated, and embedded formats, letting anyone provide the directed sports content they need for a website or article. According to founder Robbie Allen’s latest blog posts, the statistics company is going to develop an artificial sports journalist, one that will take college basketball data and compose an original sports story. Allen aims to make this unnamed project capable of writing so that 90% of readers would be unable to distinguish it from a human journalist. He also has identified 21 different types of sports stories that could be automated, and plans on expanding the program from NCAA basketball to the NFL, NBA, NHL, MLB and other leagues. While Allen has yet to provide examples of the artificial journalism, he did tell ReadWriteWeb that the program should come online this summer. Looks like the writer’s guild is about to get another digital member.
We’ve already seen some interesting samples of automated sports writing from Northwestern University’s Stats Monkey program. Recently, we’ve also seen a Japanese robot that will perform basic photography and interviews and write a report on what it finds. Now with these announcements from Statsheets, it seems like artificial journalism is getting ready to come into its own. It makes sense for artificial writing to start in the sports arena where so much of the reporting centers on quantifiable figures. Eventually, though, we should expect robotic reporters and and software writers to expand into all fields where stories typically follow a formula.
With nothing more than an announcement to go on, it’s hard to get an idea how good the Statsheet artificial writer will be, or even if Allen will manage to create one at all. Still, if we look at the Statsheet network, we see a company that has built itself up into a useful tool in a very short amount of time. Allen began the company in 2007, focused on college basketball statistics. Now there is an entire Statsheet network and the coverage is continuing to expand into the NBA, NFL, college football and NASCAR. Each branch of the Statsheet network focuses on a different task: Stat.us gives you Twitter-like updates, StatFix gives you quick facts about favorite teams, StatPlot lets you make graphs, and <Embed> lets you incorporate these different facets into your own website. Here’s a video explaining the latter:
Statsheet, then, conceivably has the expertise and the passion needed to venture into artificial journalism. It already takes statistics supplied by vendors and meld them into higher order evaluations of the games. The next step is simply to take those evaluations one step further and transform them into formulaic sports writing. On the other end of things, facial recognition, and video analysis software may eventually allow computers to collect the statistics as well as use them.
Anyone who worries about losing their job to these artificial sports writers, however, should know that Allen sees these tools as augmenting human writing not replacing it. In his blog his surmises that such systems will be best used (at first) covering schools and games that major news outlets don’t have the resources to cover. Paying a human to write a story about a high school game that only a 100 people will read doesn’t make much sense, but a robot can write a thousand such stories as easily as it writes one.
Take a longer look at this trend and it’s easy to worry about the job security of journalists everywhere. Just as we have chatbots learning to talk like humans, and robots learning to draw like humans, we’ll eventually have software that can write like a human. It’s too soon to know if the Statsheet journalist will prove an innovative step towards that goal (or if it will ever be made), but even if Allen’s project fails, another will rise to succeed in its place. Robots may never be able to top Shakespeare, but Saenz?…that dude better watch his back.
[image credit: FuturePerfectPublishing]
[source: StatSheet Blog, ReadWriteWeb]
http://statsheet.com/blog/automated-sports-content-the-future-of-sports-journalism
