Personalization, Filter Bubbles, and civic duty

It was late last year that I first described my apprehensions in regards to Google Social Search. I maintained then, as I do now, that Social Search facilitates motivated reasoning—a phenomena I tried to explain in relation to Mercier & Sperber’s Argumentative Theory of Reasoning. According to this theory, “reason” isn’t the logical and truth-oriented thought process it’s made out to be. Quite the opposite. Mercier & Sperber suggests that reason has as its primary goal to help us win arguments which, if they are right, means that reason is motivated by preconceived ideas.

I will elaborate on this idea in the next couple of posts. And in this first iteration, I will focus on two ideas that has helped shaped my thinking around this issue since I first wrote about it back in August.

Idea #1 – Filter bubbles

The first and, perhaps, most interesting development is the publication to Eli Pariser’s book “The Filter Bubble”. Here, Pariser raises concerns very similar to my own, albeit it with one very important distinction: he doesn’t limit his criticism to Social Search or, for that matter, Google. He writes:

Sites from Google and Facebook to Yahoo News and the New York Times are now increasingly personalized—based on your web history, they filter information to show you the stuff they think you want to see. That can be very different from what everyone else sees‚ or from what [you] need to see.

Pariser’s concern is that personalization envelops users in bubbles of targeted content and that these bubbles insulate us from new ideas and experiences. It’s sort of like an Echo Chamber but not quite. In Pariser’s filter bubble, you don’t need to say anything. Your actions — the pages you visit, the links you click — speak for you, providing Google et al with the information they need to second guess your likes and dislikes.

Idea #2 – Critisism

The second thing I want to mention is the downpour of critizism ventured against Pariser and the notion that that personalization is in any way a bad thing. John Hawks isn’t alone in pointing out that “At no time in history have people been exposed to a wider range of opposing viewpoints” which, I have to say, is hard to deny. He writes:

We have always had algorithms to select content. In the past, those algorithms were inside the heads of a small number of newspaper editors and media programming executives…That’s why you see things like different newspapers, owned by different companies, publishing opinion pieces on the same out-of-the-blue internet theme on the same day! It’s like a throwback to the past. I like Google better.

Clearly, we don’t need the Internet to construct filter bubbles for us. We’re more than capable of doing it ourselves or, as in Hawks’ example, get other people to do it for us, either by reading particular magazines or watching news networks with a certain ideological bent.

A call for civic duty

Hawks makes a good case, I think, and Pariser must have anticipated it because he goes on to note that newspaper editors and programming executives have a sense of civic duty. They are not algorithms that filter content indiscriminately. People and algorithms may both serve you, the user, with a fair amount of LOLCats if that’s what you’re into (hey, what ever works for you!) but the person, Pariser suggest, would at some point make sure you also get coverage of the war in Iraq, say, or the recent outbreak of E. coli in Europe.

And let’s be fair here: Pariser isn’t arguing for personalization to go away. That, I suspect, would be futile. What he wants instead is for the Googles of the world to infuse into their algorithms that sense of civic duty. He wants the algorithms to behave a bit more like those newspaper editors that says “OK, that’s enough LOLCats, now look at this in-depth analysis of the Romanian elections instead.” (Google already does this to a certain extent but the details are murky).

This is the point in Pariser’s argument where I must raise my hand and politely disagree. Because whose sense of civic duty are we talking about here? Are we talking about Google’s sense of civic duty? Or maybe it’s up to the engineers at Yahoo!? Or maybe Facebook gets our vote? I don’t know about you but I’ve been around for long enough to know that the sense of right or wrong—not to mention what is and is not important—varies from person to person (not to mention from corporation to corporation).

Where would we draw the line and, more importantly, who’s going to be in charge of the line drawing? Politized search appears to me a slippery slope leading to more questions than answers.

The observant reader will now say that “Yes, sure, but aren’t we being a tad unfair here? It’s not like those newspaper editors are epitome of objectivity either! Why should the Googles of the world be held against a different standard?”

The badge factor

I’ll tell you why: because Hawks’ newspaper editors wear their political affiliations like a badge. No Democrat in their right mind would watch Fox News. No Republican in their right mind would read NYT. Content publishers have received tacit consent from their users to apply their own sense of civic duty and political bias to whatever content they publish. In my book, it isn’t too much of a stretch to say that this consent should encompass personalization as well. We are, after all, just getting more of what we’ve already asked for.

But the same cannot be said for search engines! They wear no badges, political or otherwise. And so we have to concede that search engines are different from content publishers. Democrats and Republicans alike use Google. They wouldn’t pick Bing over Yahoo! for political reasons simply because Bing and Yahoo! doesn’t take a political stance.

What if search engines were divided along ideological lines just like newspapers are today? Google for Liberals? Bing for Conservatives? Yahoo! for Libertarians? Isn’t that what would happen if and when users realized that their favorite search engine was trying to impose on them its specific brand of civic duty? People would side with the service they thought was most like them ideologically, and the service providers would, I’m sure, play along as long as there was business to be made.

Thanks, but no thanks

That’s a pretty bleak picture and so I’m going to have to side with Hawks on this one. I too like Google better. The non-partisan Google, that is. Politicized search clearly isn’t the way forward.

This is not to say that I’m now convinced personalization isn’t a mixed blessing. I still think it can be depending on the circumstances, and this is something that I’ll have reason to return to in my next post. For now, let me just leave you with a question: if Mercier & Sperber are right in saying that reason is motivated by preconceived ideas, wouldn’t it be both easier and more desirable to let users choose when to receive personalized content and when not to? After all, who better to determine when personalization is desired than the actual person?