Thursday, February 12, 2015

how to burst the filter bubble

Review of The Filter Bubble
by Eli Pariser
Penguin 2011

Last year, when the publication of a psychology experiment conducted on facebook highlighted the dubious ways in which the company manipulates people’s timelines, some clever observers remarked that this wasn’t really new, and it was all described in Eli Pariser’s book. Which, amazingly, was the first time I heard of the book.

Having picked up a copy for £ 2 on the flea market (somebody clearly thought it was no longer relevant) I’ve now read the book. Essentially it confirms my hunch that the internet took a turn for the worse when the major global monopolists amazon, google and facebook personalised their services and started to pretend that they know which books I want to read, which search results I want to see and which ads I might be interested in. The resulting limitation to what one can see of the internet is the phenomenon that Pariser calls the filter bubble.

Pariser explains very eloquently why this is bad news. It’s bad for creativity, because it removes the possibility of serendipitous findings and unexpected connections (which are basically the things that pay my mortgage, so bad news for me especially!). It’s bad for democracy, because people only surrounded by the opinions they agree with and the friends that have similar views will fail to see that there may be another side to the argument. The anti-Islam movement Pegida in Dresden and other German cities originated as a Facebook group –although long after the Filter Bubble was published. And far from fulfilling the internet’s early promise of distributing power to the people, the application of opaque filtering methods concentrates unseen and unaccountable power in the hands of a small number of global companies.

What can we do about it? One can’t completely opt out of it, as an increasing number of services would just stop working if you chose to block all cookies. Pariser says companies should be open about what kind of filtering and personalisation they apply to any given product. But seeing that the key players hold quasi monopolies, what are the chances that they will comply with Pariser’s wish? Four years after his book was published, I haven’t seen it happening, and the affair around the Facebook experiment seems to suggest they are still blissfully unaware of the problem. Governments could perhaps legislate to enforce more openness, but then again, they never understood the internet anyway, and are already failing to make the companies pay their taxes. Not much hope there either.

Which leaves only consumer choice. I have always preferred social sites with an element of anarchy and randomness to facebook with its emphasis on real names and real friendships. Pariser, although he admits to enjoying facebook as much as the next person, praises twitter for its unfiltered timeline and clarity regarding where filtering is applied. For instance, in searching for hash tags, you can choose between “top results” and “all results”. I would add tumblr to that recommendation, as it also has an unfiltered timeline and the most amazing random people on the planet. (As far as I know, my 1260 followers on tumblr include only one person I have met in real life.) And if you do use Facebook (as I grudgingly do once a week), make an effort to venture outside your bubble. Befriend random people, do unpredictable things. Confuse the algorithms.

Similarly for search, try to think beyond Google, maybe use several search engines. I use duck duck go as default, and only resort to versions of Google when I need results in a different language, so for instance, google.de for family history explorations in Germany. Google’s bots must think that I’m only interested in dead people.

As for amazon, as I’ve explained before, they haven’t managed to figure me out in 15 years, and I don’t think they ever will. It would take a human being to actually read what I write, and they only have computers. So my remedy is: be wide ranging and unpredictable. Surprise yourself and the algorithms. Burst the filter bubble.

No comments: