lunes, 21 de enero de 2013

Facebook Graph Search Is Humorless, Creepy And Doomed To Disappoint

Facebook's newest feature Graph Search (so new it's still in beta) can apparently tell you lots of stuff. Which of your friends are into surfing, hiking or drinking cups of tea. Or, delving into darker territory, which of your friends are sexist or racist — as this Gizmodo article points out. (Ewww.) But Graph Search also shines a massive illuminating spotlight on something else: why Snapchat has captured people's imagination. And the answer is simple: because it does something Facebook does not. It lets data disappear, rather than stockpiling it until it starts to stink.

Snapchat is a photo-messaging app that lets you send a missive that automatically deletes itself after a set amount of time (from one to 10 seconds). Which means your Snapchat buddies' proclivities and peccadilloes linger mostly in your memory (and any screengrabs/photos you've been able to take – but that's a whole other story). Snapchat's existence — and apparent popularity — is fascinating in an age of indelible digital footprints. YOU may not remember that in 2007 you (jokily) liked 'Barbara Cartland novels' on Facebook but Facebook sure as hell does. And now that we have Graph Search lots of other folks can know something about you that isn't true.

Facebook's algorithm isn't actually that clever, so it can't figure out that you were being ironic when you clicked the like button. Leave your humor at the door, all who enter here. (As an aside, there is a huge festering issue for Facebook and its advertisers in the form of so-called 'dirty data' – the site is awash with people playing deliberately fast and loose with the truth. Teenage girls who are 'married' to their BFFs. People who work at 'Rydell High' or 'Charlie's Chocolate Factory'. Or who look an awful lot like Brat Pitt. And that's before you even get to the whole fake likes issue – explored in some detail here. In short: just because it's called a 'like' doesn't mean someone actually likes it.)

The rise of Snapchat — Google Play lists its Android app as having clocked up between one to five million installations in the past 30 days which, while no Angry Birds, isn't trivial either — runs counter to claims that young people don't give a damn about their personal data being perpetually published online. We don't yet have detailed demographic data on Snapchat users but there's plenty of anecdotal evidence its use is being driven by teens and young people (13 to 25 year olds). In other words exactly the group who were supposed to have been born post-privacy. Don't R.I.P. privacy – we're coming with a shovel to dig you back up.

The surprise success of an app that deletes rather than retains data suggests there is potential for other startups to get themselves noticed by letting data decay. By building more transient, human technology that seeks to mimic the privacy of memories – which remain, for now at least, locked safely in our skulls. Evidently there is an appetite for digital footprints that – like a real footprint – gently fade and crumble away, leaving little or no trace of your passing. After all, even Facebook decided it needed to clone Snapchat – with Poke. So let me say that again: startups, there's an opportunity here. Think about ways to humanise technology. To make it a little bit more forgetful, a little less 1984. Seize the day. And you too might have Zuck and Co. running scared.

Technology that's a little more human, a little less stalkerish could have a very bright future, indeed. Especially as more and more companies start to join the dots of people's personal data – illuminating the implications of continuing to contribute to a single catch-all data bank. What is actually in it for the user? Not necessarily much of value when that data is dubious or paranoid or both. I explained Graph Search to a twentysomething U.K. Facebook user who grew up using the service as a teenager. Her response was a question: "How does this Graph Search benefit the people who use Facebook?" How indeed. If you aren't already aware that your surf board-owning friend likes surfing, or your travel-mad friend went to New York City for New Year's Eve, are you sure you're actually friends with them?

After Facebook unveiled Graph Search, my colleague Josh Constine argued that not sharing your likes on Facebook is now a selfish act because you're depriving humanity of your own humble contribution, insights, knowledge, tastes and so on. In my view the exact opposite is true: by sharing anything on Facebook you risk over-sharing more than ever – which means that Facebook risks making more of its users think long and hard about every little thing they contribute to its data banks. The visibility of Graph Search contributions can apparently be controlled in the privacy settings, but as is always the case with Facebook and privacy, it's not immediately obvious who will be able to see what — meaning you have to put the leg work in figuring this stuff out on Facebook's behalf. And that's just annoying. This uber joined-up social network is hacking your life back together without asking – and probably making a hash of it (i.e. you) in the process.

Tools such as Graph Search – and more broadly, whole social networks designed to make it easy to browse through large chunks of personal data, chronologically and/or by slicing and dicing it in various ways, while simultaneously making it insanely complex to opt out of putting personal stuff in the public domain (networks which take individual bits of data and sew them together to make a Franken-YOU) – are playing with fire. And promising far more than they can realistically deliver.

By contrast, one of the best things about Twitter is that it knows the importance of limits: 140 characters per post, and a site structure designed to take time and effort to go delving back into a user's catalogue of tweets adds up to a service that feels more respectful of its users.

Cumulative, joined-up data is simply not the same beast as data that's distributed and at least partially disconnected. And by pretending that joining up all those disconnected dots is no big deal, Facebook is being dishonest and dumb. Plenty of people won't care what quasi-identity Facebook gives them, but there will be others who decide there's no benefit to them of being on Facebook's public record. Indeed, that removing their contribution is the most sensible response to its behaviour. In real life only stalkers and psychos keep databases on their friends.

Share out pieces of a jigsaw puzzle between your friends, family, lovers and strangers and it would need a freak accident to gather the whole odd-ball collection in one place to put the picture together. But even that metaphor is way too simplistic. The point is: there is no one you. And any algorithm that tries to create one by joining your disparate dots is pitifully reductive – and doomed to fail by overpromising, under delivering and being really really creepy while it flounders around trying to impress you.


No hay comentarios:

Publicar un comentario