The word ‘googling' has, in our Internet age, become a somewhat acceptable verb for searching for information on the World Wide Web. The ‘naïve' Web users that we are, we believe that the ‘10 blue links' that pop up in response to our queries are fairly sacrosanct; that is, your search results are no different from mine.
However, critics allude to a growing information dystopia on the Web. Most famously, at a TED talk this year that went viral on the Web — where else — Eli Pariser spoke about what he calls the online “filter bubble”. In his recent book, titled The Filter Bubble: What the Internet is Hiding from You, Mr. Pariser makes a compelling argument against search personalisation, a strategy that is widely believed to be the future of Web search.
Web personalisation is fairly ubiquitous. For instance, micro-blogging site Twitter offers a personalised recommendation service, both for who you may want to follow and what you would like to see. But Pariser's observation is far more important, and disturbing. He uses a compelling example to present the invisible shift in information flow on the Web by comparing search results encountered by two friends.
When they searched for the phrase ‘Egypt' (at the height of the Egypt revolution), one found links to the political crisis and protests in Tahrir Square, while the other found his search page full of tourist information on the country.
This invisible, algorithmic editing of what you see on the Web is problematic, critics have pointed out. So the idea of a standard or objective search could be entirely notional. Even if you're ‘safely' logged out, as Mr. Pariser points out on his blog, your search is being tailored. Internet search firms offer a ‘personalised' (or customised) version of what the system — driven by sophisticated algorithms that tap into your search history, previous clicks and Web services you use — believes you would like to read. The algorithmic technology, that we believe is used to target only advertisements, is now also extended to content.
Viewed in a political context or in the context of information flow, indeed personalisation could end up blinkering your ‘world-view'. News organisations too, he believes, are investing in this technology. Mr. Pariser calls this “auto-propaganda”.
Indeed, there is an information overload on the Internet; we encounter this every time we seek specifics on the Web. The growth of social media and user-generated content has lead to an information explosion. But are these filters, controlled by Internet firms, the solution?
Internet firms particularly ones offering multiple services such as search, email, social networking and so on, are privy to a lot of user information. Privacy concerns apart, the arguments Mr. Pariser puts forth in his book primarily revolve around the socio-political implications of creating such bubbles.
How does this work? Mr. Pariser claims that Google uses 57 signals to customise. Your IP location, the browser you use and your search history is used as material by the algorithm (tech parlance for a set of computer instructions) to customise your results page. The algorithm takes all this information, extrapolates it and attempts to predict what you're most likely to want to read and what interests you.
Indeed, this means that your 10 blue links are quite likely to be different from mine, as Mr. Pariser's examples illustrate. The ‘Internet bubble' is very visible on social media. While a random comparison of search results may not yield hugely visible difference, what is very visible is that on Facebook, your feed automatically edits out friends that you don't interact with as much, or whose walls you don't frequent.
Pariser's critics feel that personalisation may not be as evil. After all, it is great for consumers and also makes business sense (the better they serve you, the more ads they draw). And to be fair, Google does offer a ‘turn off customisation' button, albeit barely visible and tucked away at the bottom of your search page.
In a recent interview with The Hindu , a Yahoo! search expert said that despite customisation, increasing diversity is very much part of search strategy.
