The Application and Management of Personal Electronic Information

Recently the First International Forum on the Application and Management of Personal Electronic Information, organized by the MIT SENSEable City Lab, gathered many stakeholders from multiple disciplines to share on the issues surrounding the application and management of personal electronic information:

The goal of this forum is to explore the novel applications for electronic data and address the risks, concerns, and consumer opinions associated with the use of this data. In addition, it will include discussions on techniques and standards for both protecting and extracting value from this information from several points of view: what techniques and standards currently exist, and what are their strengths and limitations? What holistic approaches to protecting and extracting value from data would we take if we were given a blank slate?

Luckily, many of the position papers and presentations are now online.

Several contributions look at other field such as health care to draw best practices of personal logs storage and mining. Particularly, in Engineering a Common Good: Fair Use of Aggregated, Anonymized Behavioral Data, Nathan Eagle argues for the necessity of a set of standardized protocols for behavioral data acquisition and usage to preserve both individual privacy and value of the community. Nathan has been analyzing behavioral data from mobile phone operators to help epidemiologists modeling human movement to support the allocation of malaria eradication resources in Kenya. With similar data, he supported planners of Kigali in quantifying the dynamics of slums and the social impact of previous policy decisions ranging from road construction to the placement of latrines (see Artificial Intelligence for Development). Still there are two major issues in the use of these data even if anonymized and aggregated:

• Deductive disclosure: the nature of behavioral data is such that very few observations are required to deduce the identity of an individual. An issue that is overcome to some extent by strict data sharing protocols that ensure the data cannot be released to the general public. Other strategies my apply to some extend as well (see On Locational Privacy, and How to Avoid Losing it Forever and Jon Reades’ Using Finite State Machines to preserve privacy while data mining the cellular phone network)
• Data retention and erasure: the inability of individuals to remove their data from these aggregate datasets. Good practices can be gained from the medical community that pushes for legislation enabling individuals to own their personal health records to prevent this type of exploitation. Similarly, there is also pressure for legislation on the ownership of personal behavioral data, providing individuals with the right to access and remove their data from corporate databases enabling them to ’opt-out’ from any type of analysis. This leave me wondering to what extend the opt-out impairs the quality of tha data?

Despite the necessity of rigorous data-sharing protocols, Eagle also considers of intellectual property of data can be considered as a form of intellectual property.

The behavioral IP of an individual should be owned by that individual, and licensed to third-parties for a fee if desired. The behavioral IP of a society should be considered as a valuable public good.

This certainly opens new interrogations on the applicability of this proposal (e.g. who determines the fees, who has access for free and who does not?, how to finance the efforts that transform data into valuable public good? are the developed algorithms also a public good?). In addition to discussing the IP of the data, I often argue on the necessity to apply transparent processes in which everybody is aware on the mechanisms to generate the information (see my World Information City Doggie Bag).

On that very aspect of data process (and its transparency), I was intrigued by Trevor Hughes’ (Executive Director, International Association of Privacy Professionals) intervention on “Data Environmentalism” that argues that we should focus less on “notice and choice (fair information practices) and actually put our efforts in in securing data, data flows, and legitimate use, to the point of developing Indicators of trust Transparency.

Another aspect on the exploitation of personal electronic information lies around the notion of dream of the perfect technology and the myth of the perfect power (see Stephen Graham at World Information City). It is one of the theme that Aguiton et al. cover in their contribution Living Maps: New Data, New Uses, New Problems quoting Bruno Latour in Paris: Invisible City:

Megalomaniacs confuse the map and the territory and think they can dominate all of Paris just because they do, indeed, have all of Paris before their eyes. Paranoiacs confuse the territory and the map and think they are dominated, observed, watched, just because a blind person absent-mindedly looks at some obscure signs in a four-by-eight meter room in a secret place.

On the application front per se, it is very well worth checking the recent research of Skyhook Wireless on their own data (Aggregated Location Requests) to perform time/space based analysis, frequency/phase domain extraction and baseline/anomaly detection.