What if you could pick up a printed newspaper, but instead of a handful of stories hand-picked by a secret cabal of senior editors in a dingy newsroom somewhere, it had pieces that were selected based on what was being shared — either by your social network or by users of Facebook, Twitter etc. as a whole? Would you read it? More importantly, would you pay for it?
You can’t buy one of those yet, but The Guardian (see disclosure below) is bringing an experimental print version it has been working on to the United States for the first time: a printed paper that is generated entirely — or almost entirely — by algorithms based on social-sharing activity and other user behavior by the paper’s readers. Is this a glimpse into the future of newspapers?
According to Digiday, the Guardian‘s offering — known as #Open001 — is being rolled out later this week. But you won’t be able to pick one up at the corner store: only 5,000 copies will be printed each month, and they are going to the offices of media and ad agencies. In other words, it’s as much a marketing effort at this point for the Guardian (which isn’t printed in the U.S.) as it is a publishing experiment.
Robots write news — why not edit it as well?
As tiny an experiment as it is, however, the Guardian project raises some interesting questions. A paper produced by robots (or at least algorithms) isn’t all that different from tools like Paper.li or even the “Most Shared” feature that many newspapers have now on their websites. Sharing-analytics company NewsWhip recently put together a look at what front pages might look like if they were based on what people actually shared. But is that what we want from a newspaper?
The Guardian‘s latest project is based on a similar experiment it has been running for the past six months or so in Britain, one that generates a printed paper called “The Long Good Read,” made up of some of the best long-form content from the Guardian and its sister paper The Observer. It’s available for free once a week at the Guardian‘s public coffee shop in the London neighborhood of Shoreditch (a shop that is an interesting experiment in itself, as I’ve discussed before).
The Long Good Read started as a joint venture with The Newspaper Club, a company that prints small-run custom newspapers, and was based on work done by former Guardian developer Dan Catt as a side project — a way of automatically collecting the best reads from the paper for later reading, as either an RSS feed or a sharing feature similar to Longreads.
As the editor behind Long Good Reads explained in a blog post, the paper uses algorithms — including the Guardian‘s own in-house tool for tracking which stories are the most read and the most shared — and generates a list, which The Newspaper Club robot lays out in newspaper style. All the human editor does is check to see if any stories are out of date by the time it gets to the printing stage, and/or fiddle with the layout a bit. The whole process takes about an hour.
“Sometimes what the Guardian decides to put on its front page or home page matches what the users are reading, other times everyone seems to be focusing on something else, with twitter, facebook and other sites acting as a back channel. Editors are our first filter, the readers are our second.”
Will we miss the serendipity factor?
Obviously, neither the Long Good Reads project nor #Open001 are going to replace The Guardian or any other printed newspaper any time soon. One is a monthly publication and the other is weekly, and they are aimed at what — for now, at least — are fairly niche markets. Nevertheless, we clearly have the ability to print newspapers tailored to our interests. But should we?
The benefit of tools like Paper.li — which allows Twitter users to create a custom digital “newspaper” view of all the links that have been shared by people they follow — or even customized digital magazines like the ones Flipboard launched last year, or Facebook’s new Paper app, is that we can create a specialized news-feed that is targeted directly at our interests, via our social graph.
One of the downsides of this approach, however, as explained by Filter Bubble author Eli Pariser (who is now, somewhat ironically, a co-founder of the viral-content engine Upworthy) is that we potentially become surrounded by things we already agree with, instead of being challenged or exposed to different ideas.
This is why news-recommendation engines like Prismatic try to engineer what they call “serendipity” features, so that everything you see isn’t a homogeneous mass of things you have already expressed an interest in. Newspaper editors also (theoretically at least) choose to print stories that might not be sexy or interesting, but are important in some way. Can we teach robots how to do that too?
Guardian News & Media is an investor in the parent company of Gigaom. Post and photo thumbnails courtesy of Thinkstock / Ociacia as well as NewsWhip and The Long Good Read