January 14, 2012

In Defense (Ouch!) of Facebook's Data Deal with Politico

Regular readers know that I am not a fan of Facebook. I've long been (and continue to be) a critic of their privacy practices and related implementations. I don't actively participate in Facebook, other than with an essentially empty account providing me with access to inspect the various user interfaces and interactions.

So it's with a bit of irony that I find myself in the position of having to come out in support of Facebook, in response to a torrent of complaints about their initiative to analyze the political sentiments of Facebook users in conjunction with Politico.

The project involves an automated system to scan Facebook posts and comments for politically-related text, analyze the resulting data, and derive an overall look at Facebook users' sentiments on this topic, which given the size of Facebook's user community could potentially provide important new insights.

But I've been inundated with people upset that Facebook is "reading their private messages" and so supposedly horribly violating their privacy.

I'm sorry, I just don't see it that way.

I have a maxim that I've used as a guide for many, many years. It's very simple.

"Choose your battles carefully."

I don't claim to have the military insight of Sun Tzu, but this simple idea has served me fairly well.

And in this particular Facebook case, we shouldn't even be thinking of a battle at all, because there is no demonstrable vector for damage to any Facebook users through the anonymous, automated statistical process as described, and there is the potential for significant knowledge benefits to society from the results.

We're at a crossroads where it is now obvious that simply throwing away the vast and rapidly increasing amount of data being generated by our lives, without first considering the responsible derivation of useful knowledge from that data, is like flushing a natural resource down the toilet.

Obviously, there is much data that shouldn't be kept in its raw form forever, to help protect against the possibility of retroactive abuse by governments or others. That's why I've applauded the now fairly standard practice of scheduled data anonymization and deletion by major search engines and other services for user search query and other data, for example.

On the other hand, to argue that it should be forbidden to derive useful knowledge from data in ways that protect individual users' privacy but still allow for aggregate analysis, would be incredibly wasteful.

A classic example is health care. We all want our individual health-related activities to be kept private. But without some mechanism for the aggregate sharing and analysis of healthcare data, it would be impossible to reasonably recognize and understand trends in disease, diagnostics, and treatment. Epidemics and pandemics would be far harder to spot early on, and to potentially control. Real lives are at stake.

In our everyday world, all manner of activity data is subjected to various forms of aggregated analysis and often publication. Phone companies and ISPs watch calling and traffic. Credit card companies and banks explore transaction patterns in great detail, as do stores with "loyalty card" systems.

Internet messages are scanned by automated systems for anti-spam, anti-phishing purposes, and for ad displays. Search trends are derived from user search queries.

And so on. There is usually nothing nightmarish about these systems, so long as individual users' data is protected. And the benefits of the knowledge derived from such data can be very important. Not only do they serve to keep even incredibly sophisticated Internet services free to most users, but the health, safety, and other realms of aggregate statistical information can serve critical public needs.

So like I said, choose your battles carefully. There are enough true privacy problems to keep us busy into the dim future. Wasting time, effort, and emotion on useful anonymous data analysis like Facebook's Politico project just doesn't make sense.

Even if you don't like Facebook.


Posted by Lauren at January 14, 2012 02:39 PM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein