Path Dependence, Technology, and Privacy

-- By DanielShiner

What version of a technology becomes dominant first can have massive ramifications for privacy and freedom. Path dependence, often driven by network effects, can mean that the technology first widely adopted for a given purpose is the technology we may be stuck with until further technological developments overshadow the old.

The issue of path dependence can be seen clearly in the results of the videotape format wars starting in the 1970’s, in which various technologies were competing with each other for primacy. While many believed Betamax to be the better technology, VHS ended up winning for other reasons. The outcome of the format wars was that VHS became the unchallenged dominant choice in consumer video technology for decades, regardless of its weaknesses, until it was finally toppled by DVD.

What we can learn from the format wars is that the best technology is not necessarily the one that becomes dominant. This is true of privacy protection in social networks as well.

Once a social network gains wide adoption in a given sphere, whether it be geographic or purpose based, network effects give that social network a near monopoly in that area. This is because the primary value of a social network is not the user experience, but the number of other people you want to interact with who already use the system.

In the United States and a large portion of the world, Facebook has become the dominant social network. This dominance is primarily derived from its network effects, making it extremely difficult to challenge on its own ground. Often using huge amounts of resources to build a superior user experience, alternatives like Google+ have tried to dethrone Facebook. All of these competitors have failed, largely because they are worthless if they have not achieved the critical mass of people necessary for a social network to have value. Even if Google’s interface had been better in every way (which it arguably was not), using it was pointless because there was no one using it already.

From the perspective of privacy, Facebook's dominance has been a disaster. Facebook was built with a centralized architecture such that its users’ data is fully entrusted to it. Because Facebook has control over its treasure trove of user data, it becomes an inevitable target for state surveillance. What we are left with is a dominant social network, whose existence has become essential to the surveillance schemes of various governments around the world. In effect, the way Facebook was built made it destined to become a quasi-governmental surveillance tool.

The seemingly impenetrable fortress created by Facebook’s network driven monopoly creates a problem for people who value both social networking and privacy. If I am uncomfortable with Facebook’s centralized architecture, and prefer a more distributed and transparent social network in which I control my own data, my options are not good. First, I can stop using social networking all together, but if I value social networking the cost of this could be prohibitively high. Second, I could leave for an alternative social network like Diaspora, but the whole point of a social network is that it is “social” and if the people I want to interact with are not already on their system, I am in effect quitting social networking all together.

No. You didn't bother to learn about Diaspora, which was designed to solve that problem. Undermining the network effect you are talking about is comparatively simple. You can describe why it won't work in your considered opinion once you understand it: that's fair argument. But ignoring away the whole point of other peoples' thinking, whether it is expressed in words or in software, is not intellectually reponsible.

Third, I could create my own alternative, but will face the same problems that established alternatives have, which is the fact that Facebook’s value stems from the fact that so many people are already on their system. Fourth, I could complain to Facebook, but because they know my alternatives are bad, they are unlikely to make meaningful changes.

In the end, we are left with what is effectively an all or nothing choice. For now, it’s either Facebook or returning to a pre-social networking age. While people who value privacy above all else will find this an easy choice, Facebook’s popularity in the post-Snowden era illustrates that the choice is not as easy for others to make.

What path dependence means for the future

The lesson of Facebook’s domination is an important one that privacy advocates should internalize and learn from. Ex-post reactions to new technological developments are not enough in cases where path dependence is at play. This creates a difficult challenge in which the protection of privacy involves actively creating new technology crafted to protect its users. But if someone succeeds in this, path dependence can help lock in privacy enhancing technologies against competition from even well funded competitors.

Using path dependence to protect privacy will continue to grow more important over time as future technologies with even more privacy destroying potential become feasible. Augmented reality and direct neural interfacing are two related areas in which path dependence could alter the trajectory of humanity. If the adopted versions of these technologies are built in a transparent way that allows users to keep their own data, it is possible for decades of challenges to be averted. While it verges on science fiction, if the dominant version of these technologies is built to centralize information and deliberately makes its operation opaque (like an Apple product), even privacy of thought may become a thing of the past.

To the extent that certain technologies are inherently path dependent, the future of privacy will depend on getting privacy protecting versions of those technologies adopted before worse alternatives have a chance to do so. This is a monumental challenge, and may be impossible to orchestrate, requiring that those on the frontiers of innovation value privacy and are committed to make sacrifices in order that their creations reflect their values. With user data being such a valuable commodity in today’s world, this commitment is not one that will be easy for many to keep, but the future of privacy may depend on it.

This is not a satisfactory form of the argument for two reasons: First, as I noted above, it's technically wrong. Looked at more closely, the network-effects path dependence of social networking is a marketing claim, and one which neither the data about the world nor the data about what the people running the networks really believe bears out. Twitter's and Facebook's history as firms both show why the "stickiness" of social networks is quite low, and the stickiness of social networks that earn their profits as ad platforms running auctions for access to eyeballs is even lower. Facebook's WhatsApp? strategy, like its "free basics" system for data-mining the packet traffic of the poor show that it knows that its own arguments about the stickiness of its network are false: it's moving as fast as it can to the post-Facebook future before the herd moves elsewhere.

Second, more generally, this form of invoking "path dependency" is an extremely cut-rate form of simple historical determinism. The future is a foreseeably-determinate collection of path dependent outcomes, which happen to add up to the indefinite prolongation of present phenomena. This is not quite as elegant as "whatever is, is right," but it is both as socially disarming, though less normatively obvious, and as completely historically false. Path dependence and complete contingency are entirely compatible, which is the pervading argument of Steven Jay Gould's masterpiece, "Wonderful Life."

It seems to me that improvement can be realized in both directions within the scope of the present inquiry. If one backs away from the use of path-dependence as denial of contingency, restoring it to its real role as a principle of historiography—that the past is not completely but is firmly prologue to the present, and direct historical analogy is therefore a weak form of argument—one can begin to ask what does shape the technological precession of modes of social communication in a situation as rapidly-changing and flexible as the present. Obviously incumbents do what they can to make incumbency itself a shaping force, investing in the creation of inertia in the guise of "product improvement" or, more generally, "innovation." But actual innovation also occurs, and sometimes for social purposes. Given the history of recent changes in the way humans make software, one might even come to the conclusion that the latter form of genuine innovation is rather powerful. How does that interact with the force of incumbency? If we really do live in your form of path dependent world, then one might be tempted to conclude that what happened to IBM, DEC and Microsoft in the last three eras of computing will necessarily happen to Facebook in this one. But that's apparently a bridge too far for you.