Wednesday, November 09, 2011

The anti-social nature of social networks

Whilst working on other things I came across two excellent essays on social networking today, one from maciej on Pinboard and the other from Charlie Stross. The former articulately describes the highly anti-social nature of the architecture of social network services. The latter critically assesses a social network analytics firm.

There are echoes of Alessandro Acquisti's SCL 6th Annual Policy Forum keynote address in both. Acquisti is interested in the behavioural economics of privacy.

Maintaining privacy in a social networking context involves engaging in the complex calculus of balancing an immediate cost (eg deciding not to participate) against a future benefit (preservation of privacy) and we're not particularly good, as social creatures, of tipping the balance in favour of the long term benefit.  Hence the traditional Schmidt/Zukerberg 'solutions' to privacy -

[let the market decide]/[no one's forcing you to participate]/[transparency]/[user control] = privacy expedited

don't work.  Counterintuitively, the experiments that Acquisiti has done demonstrate that the user, who believes that they have more control over their personal data, trusts the entity offering that superficial control and ironically discloses more personal data. Whereas natural suspicion of economic agents (companies) not offering that apparent control leads to more limited disclosure.

One of the many other problems with mass personal data pollution is that the negative events/perceptions show a greater longevity than the positive.  Or as Acquisti says, when it comes to privacy, "the bad is not only stronger than the good it is also discounted differently."

If there is a record on a social network or elsewhere in the personal data morass on the internet of someone getting a major award, engaging in a heroic act or just doing something which shows them to be a decent human being, the social kudos associated with that postive event has a limited life.  If it was 5 years ago, for example, we ask what has he done in the meantime?  But a negative event is much longer lasting - a bad deed five years ago will still be held against you.  That's human nature.  The net takes away the ability to forget and the human pscyche makes it difficult to forgive, particularly in an era of a 24 hour news cycle with a voracious appetite for bad news.

(Just an aside, it's been impossible in the spotlight of that news cycle, for any kind of rational public debate to emerge on the border control story in the past few days because of the competition in shrill tough talking hysteria; and the opportunity to explain the power of intelligence led surveillance compared to irrational and unworkable blanket/mass surveillance is lost in the fear of potential media accusations of being soft on immigration/terrorism/crime/[pick your favorite bogeythreat])

  1. the architecture of social networks, 
  2. our lack of capacity to weigh up or fully grasp the myriad of complex down stream uses of our personal data (we just cannot predict as consumers/citizens how our data will be used), 
  3. our psychological tendency towards immediate gratification bias (give up data now, hang the consequences later), 
  4. the rapid devaluing of past good deeds compared to the long-lasting impact of the bad 
  5. and the power of technology to mine "anonymous" data and link it to individuals (eg through improving face recognition technologies) 
  6. before we even think about the relative economic/information power relationship disparity
mean that the proposed solutions of
  • transparency - telling the user this is your data and this is how we use it
  • and superficial user control of data (here are some privacy setting buttons)
are not an adequate response to the complex problem of ensuring respect for private and family life in an information age.