We have a tendency to think of openness as the opposite of privacy, and in particular that opening up data carries the risk of exposing our private information. You don't have to look far to know that those risks are real: that logs of taxi journeys may reveal information about the places celebrities frequent or that logs of cycle hires may reveal individuals’ travel patterns.
But, as I discuss here, openness could also be used to protect people's privacy.
Openness of the Data Sharing Process
It is unrealistic to expect that data collected by government or private companies will never be shared. Indeed, there are reasons for sharing data — to improve the services we receive, to perform research, or to detect threats — that may in some cases override our human right to privacy.
Openness about what data is shared, with whom, and for what purpose, could be used to provide a natural limitation on that sharing. Transparency about access to data could disincentivise unnecessary access as well as providing a justification and rationale for the sharing that does go on. It could help to inform debates about which data should be shared and reused. In the UK we have seen this pattern of publishing data about data access requests be used by the Department for Education as it grants access to individual-level data in the National Pupil Database.
What data should we expect those who share data to make available about what they have shared and with whom? How can we ensure that there is not only transparency but also accountability when governments and companies share private data?
Openness of the Results of Analyses
Analyses over personal data often result in statistics that could be made openly available to everyone. Opening up these results could ensure that when others want to perform the same analysis in the future, they do not have to get access to the individual-level data themselves.
Monitoring the analyses that are carried out on personal data could enable those who control access to the data to understand what they should be doing to satisfy the demands for data. This may be releasing new open datasets, or it may be providing services such as the Justice Data Lab, which creates reports about the effectiveness of interventions that aim to reduce re-offending rates.
Should access to personal data collected by the government be contingent on the public release of the results of any analysis done on that data? How can data owners and the public best learn from the analyses performed by researchers? What types of service could the holders of personal data provide to reduce the need for access to individual-level data by researchers?
Openness of Reasoning
The concrete impact of access to private information is that decisions are made about us, both individually and as a society, by those that have access to that data.
In the recent Anonymisation: Techniques, Risks and Benefits Symposium 2014 organised by the UK Anonymisation Network, Mireille Hildebrandt from Radboud University Nijmegen argued that openness about the rationale and methodology behind those decisions, and the data that was used to make them, could help ensure that the data is used fairly, as well as promoting an informed debate about the way that data is used.
Analyses that we have carried out at the Open Data Institute, such as into the impact of closing fire stations in London, illustrate that this principle of openness about decisions informed by data can apply at a policy level as much as an individual level.
Should our concerns about privacy be focused on access to data or on the way that data is used? How should organisations that make decisions based on data explain their rationale to those affected by the decisions?
Openness should be put in service of protecting our privacy rather than in opposition to it. "Sunlight is the best disinfectant;" transparency should help constrain unacceptable uses of personal data while equipping us to have informed discussions about how our data is used for everyone's benefit.
This guest blog was written by Jeni Tennison, Technical Director at Open Data Institute.