Thursday, March 29, 2018

Cambridge Analytica and the hysteria around it

Will you be ware next time?

No matter how intelligent, advanced or evolved humans become, we are all gullible, at some stage. More than being gullible, we as a species are extremely lazy, even when it comes to thinking. We simply do not want to think. We want that job to be done by someone else. Over the past week, an otherwise little known company, Cambridge Analytica (CA) has made headlines across the world. The company, in its marketing and promotion material has claimed to have influenced election results across the world, particularly in getting Trump elected as the president.

The CA wild fire was sure to reach India and it did earlier this week. The whistle blower, Christopher Wylie, claimed that the Congress and the JD(U) had engaged CA to carry out research during elections. Some reports claim that CA thorough its Indian arm was also helping the BJP. After the information started to come in on involvement of Indian political parties with CA and its Indian arm, news rooms have become high pressure volcanoes. Arnab Goswami is trending #CongSmokingGun and Times Now is busy with #DataChorDossier. NDTV reluctantly managed to tweet once on CA leaks in the past six hours. But what exactly are people outraging over?

Broadly there are two aspects of the CA leaks. One, data from Facebook users was illegally harvested and two, the data was used to “influence” voters. This is the case in US. In India CA has been active since 2003. This report mentions its assignments in India from 2003 to 2012.

In 2003 SCL, CA’s India arm carried out a “Psephological study and opinion poll” for a national party to identify swing voters in Madhya Pradesh. In the same year, in Rajasthan, SCL did an internal (party audit) and external survey for a political party to understand voter behaviour.

In 2007 a political party in UP got a party audit and a census of politically active individuals through in-depth interviews. In the same year SCL carried out a research communication campaign for countering Islamist radicalisation in six states.

In 2009 general elections SCL managed the campaigns of many candidates using their “proprietary data collection methodologies” for successful campaigns.

In 2010 they carried out a detailed research programme, in Bihar, targeting 75% of the households to identify caste and its corresponding message.

In 2011 and 2012 SCL carried out research to understand caste and its dynamics to leverage it during elections.

Internal party audits should be least of our worries. What we should look at is the work done by SCL outside the party offices. Primarily it falls into the category of field work. What I think happened was something like this - A kind of caste census was done by SCL to tag a house with a caste. Together with other information like household income, number of family members, etc. campaigns were structured and targeted messaging was done to relate to the voters. I do not think Facebook would have played a major role in identifying caste of people. And such data collection drives are not unique. They are routinely done by almost all major parties, across the country. Most of the data is anyway available in the Census reports.

Why don't you buy a ticket to
Krabi? Or Siem Reap?
The other area where Facebook data was used to “influence” voters in US, is of grave concern. The data of 50 million Facebook users was used to send them targeted ads and posts based on their psychological profile. Now this may seem scary, but we have all, at some point in time witnessed such an attempt. Let me give you an example. I was looking up for flights to Krabi and Siem Reap over the past two days. During my research for this article I came across the page of The Hindu and viola, there was an ad from Yatra, selling me flights to Krabi and Siem Reap. This is exactly what CA did, only the other way around. They profiled people and sent pro Trump ads/posts to swing voters or anti Trump voters. There would never be enough proof to suggest that such campaigns actually work.

Our online presence has always been monitored by the likes of Google, Yahoo, MSN, Facebook, Amazon, Flipkart, Makemytrip, Twitter, etc. Our browsing history, purchase patterns, online reading behaviour is all known to these tech giants. They might not know our names, but they know our IP. What CA did is nothing new. Only they put the data to influence voting decisions by sending them customised campaign ads or posts. Some people might get outraged by this act. But then let us see what the political parties have done traditionally to “influence” voters. Parties have promised laptops, TVs, free internet and electricity, religious trips and so on. The migration from, in your face bribery to more subtle online messaging is perhaps the natural course a political party will take. Especially when the young voters are getting most of their content from internet, rather than from news rooms.

Targeted messaging or posts or content is the backbone of social media. Once you start watching too many Indian folk songs on YouTube, it will start recommending you similar content after a couple of hours. Your home screen will be customised based on your watchlist. Similarly if you start reading or start responding to or start retweeting too many anti Modi tweets, Twitter will start suggesting you anti Modi handles to follow. It will be the same if you start reading anti Congress tweets, only then you will be sent anti Congress handles. What people are getting hysterical about, has been happening for a very long time. CA got into a spot only because they used illegally obtained data, not because they tried to “influence” voters.

We should not feed the hysteria, especially on social media. You never know who might be profiling you. What has to be done is to understand what is happening and demand rules that would protect us from future data thefts. India doesn’t even have a data protection law. Most of the popular social media sites do not have their servers in India and hence are not governed by Indian laws, whatever little there is.

What we also need to do is to be careful when we click on that “quiz” or the cool “game” on Facebook or download that fun “app” next time.  These are usually the way by which data is mined. But then one can never really be very careful with such things. We all have that one stupid friend on our list, who would click on anything that flashes.