"Testing the “end of privacy” hypothesis in computer-mediated communication An agent-based modelling approach", Paola Tubaro & Antonio A. Casilli, presentation at the Fondation CIGREF, Paris, Nov 14th, 2011
THEOP Testing the “end of privacy” hypothesis in computer-mediated communication An agent-based modelling approach Paola Tubaro1 Antonio A. Casilli2 1 University of Greenwich, London 2 Telecom ParisTech and EHESS, Paris Cigref, Research Workshop, 14 November 2011
THEOP Introduction OutlineOutline 1 Introduction Outline Research object 2 Research questions RQ1 RQ2 3 Modalities of conduct 4 Interim results 5 Expected results
THEOP Introduction Research objectDoes online networking erode privacy? Alleged tendency to renounce privacy for an open, connected existence; Younger generations reportedly less attentive to privacy issues; Are we approaching the “End of Privacy” as we know it?
THEOP Introduction Research objectImplications of the “End of privacy” A change in attitudes toward privacy would have widespread consequences; These would be ethical and legal, but also managerial and societal more generally.
THEOP Introduction Research objectSome factors aﬀecting attitudes towards privacy Literature has already identiﬁed factors that may determine diﬀerential attitudes to privacy; Particularly age but also gender, educational attainment, socio-economic status; Also: terms of service of diﬀerent online SNS providers; legal and institutional frameworks.
THEOP Introduction Research objectProblematising privacy However, online interactions complexify the very notion of privacy; Traditional notion based on metaphor of concentric circles of intimacy; Mono-directional notion (Brandeis): a core of sensitive data to be protected. ⇒ This notion no longer seems well adapted to interactions in a networked society.
THEOP Introduction Research objectPrivacy as a multi-directional, dynamic process Online privacy better described through multi-directional notion of privacy as regulation (Altman); Brunswik’s lens model: Individuals send signals to, and receive feedback from, the environment. ⇒ Self-disclosure accompanies adaptation to signals from the (social) environment over time.
THEOP Research questions RQ1Contexts and contents RQ1: Can privacy in online interactions be conceptualised as a multi-dimensional, adaptive process involving the combined evaluation of “sensitive contents” and “sensitive contexts”?
THEOP Research questions RQ1Contexts and contents (cont.) Individuals form online ties to others taking into account both value added and sensitivity of contents to share; Online ties bring forth formation of groups with shared traits, which establish contexts; Over time, contexts feed back on contents and on how their sensitivity is perceived; They ultimately aﬀect choices to form/delete/maintain ties.
THEOP Research questions RQ2Systemic outcomes RQ2: What will be the ﬁnal conﬁguration of the social system, in terms of degree of disclosure?
THEOP Modalities of conductAgent-based computer simulation Generate socially consistent scenarios on a computer; Compare their outcomes; To detect and assess variables coming into play within speciﬁc social processes; To identify suﬃcient conditions for a macro phenomenon to emerge from the interaction of micro behaviours. An aid to perform a thought experiment.
THEOP Modalities of conductThe logic of an agent-based model Generate an artiﬁcial population of agents in an environment; Endow them with basic rules of behaviour; Let them interact for a certain time and step aside; Observe outcomes at the system level at the end.
THEOP Modalities of conductAxelrod’s approach We build on Axelrod’s agent-based model of culture dissemination (1997); Each agent is represented by a vector of contents, each with a diﬀerent value; Agents form ties with each other, thus create contexts; Their values vary adapting to both content and context; Iterations of this process gradually modify the system.
THEOP Interim resultsFirst version simulation model First version programmed and run on NetLogo (Wilensky 1999); Two embedded notions of privacy: Gradual self-disclosure and adaptation to one’s personal network, through a feedback process; Binary on/oﬀ visibility settings. Some results already available; Improvements planned for the coming months.
THEOP Interim resultsResulting system conﬁgurations Figure: Stable conﬁgurations (20,000 time steps): (1) Small subnets, (2) Supernet.
THEOP Interim resultsTwo solutions emerge Many small subnets where contents are locked to contexts ⇒ “Elective communities” scenario. Supernet where all contents are shared by all individuals, regardless of context ⇒ Is this the “End-of-Privacy” scenario?
THEOP Interim resultsEﬀects of some varying parameters Figure: Number and size of nets, varying with connectedness and openness to diversity.
THEOP Interim resultsEvolution of privacy on/oﬀ settings Figure: Average privacy settings, varying with connectedness and openness to diversity over time.
THEOP Interim resultsFor further reﬂection It is when individuals grow more and more connected, and share more and more contents, that privacy becomes an issue again; The supposed “End of Privacy” scenario is in fact more complex than expected; Tendency to greater openness is not linear and may give rise to counter-tendencies.
THEOP Expected resultsExpected results and next steps Reﬁne model structure, parameters, and assumptions; Run new simulations with diﬀerent combinations of assumptions, settings and parameters, observe and interpret results; Disentangle the social processes underlying the observed behaviours of the system; Derive implications for policy-making.