Privacy Value Networks

This week I spoke at the HotPlanet workshop in Krakow. This event was designed to discuss “planet-scale” measurements, which I interpreted as how one might extend existing, small-scale, measurement experiments to cover the entire planet. So I spoke on some of the potential issues that one might encounter when measuring participants in different countries (paper here). There are clearly differences in regulation that might affect data collection, such as the Data Protection Act. From a PVNets perspective there may be differences in privacy conceptions: for instance the UK seems to be vehemently anti-ID cards, while we are happy to have 1.5m CCTV cameras, while the situation in the US seems to be more-or-less the opposite. Much of my research involves measurements of wireless device usage and human mobility, and this might be affected by cultural differences between countries: the iPhone is popular in the US, while Symbian devices are more popular in the EU, and so a study that used iPhone software (as is popular in some research groups) might have very different results in different countries. Genevieve Bell of Intel has an excellent paper looking at mobile phone usage in Asia — Indonesians tend to have more than one mobile phone each, while next door in Malaysia, one device will be shared amongst many family members. So a study that assumes that measurements from a single device represent a single person may lead to very different conclusions in different places.

What of the other presentations at the workshop? There was an interesting talk from Michal Piorkowski who has been analysing Nokia Sports Tracker mobility traces. The terms of service of this web application mean that users allow such usage of their mobility traces. This is quite useful from a researchers’ perspective. We have been struggling in the PVNets project with our study of Facebook, as the Terms of Service are quite restrictive. For instance we are prohibited from storing some user data for more than 24 hours. We could ignore the Terms of Service (and indeed many researchers do!), but this seems to be rather unethical, particularly for a privacy methodologies project. So we are faced with the utilitarian vs. deonotological ethical debate — is our research more important than the rights of the participants? We are siding with the participants. As a Facebook user as well as a researcher, I am quite happy with this decision.