Ethical Considerations This section explores the ethical considerations relevant to humanitarian actors’ potential use of CDR to monitor and respond to humanitarian crises and evaluates whether any practical and regulatory factors impact the risks. Ethical risks and benefits are, of course, difficult to quantify and highly contextdependent. For any given context, it is useful to consider each of the risks described in this section by evaluating the probability that each risk occurs and the severity of impact if the risk occurs. Power dynamics between humanitarian agencies, the government, and MNOs influence these two dimensions and should be considered in any risk assessment. For example, if the government has high capacity and significant power in its relationship with MNOs or a humanitarian agency, it may exert more pressure on these actors to share CDR data. Similarly, the political interests of both MNOs and the government must be examined, as these may impact MNOs’ incentives to share data and affect politicization of data and analyses. In SPIA’s review panel with displaced persons, several panelists articulated their concern that governments could use CDR to increase government credibility by purposefully appearing to act in the best interest of the public, even though the government’s underlying intentions may not align with the public interest.49 Involving persons who have been displaced in decision-making processes would therefore grant more legitimacy and accountability to a CDR project. A nuanced understanding of these political considerations is important. This section begins by outlining the implications of data management for ethical CDR use. Good data management is important in and of itself, and can mitigate other risks that may arise. Next, the section explores the risks of CDR use that may directly and indirectly affect displaced populations. Finally, it examines ethical issues related to informing displaced persons and the greater public about CDR use. The ethical issues covered herein are not exhaustive; rather, this section highlights ethical concerns that 26
surfaced in multiple case studies or were cited by multiple stakeholders. For many of these issues, there are pathways to minimize risks and maximize opportunities. Data management considerations
Ethical data management is essential at every step, including data collection, security, analysis, retention, and publication. Data management practices facilitate or exacerbate many of the ethical risks discussed in this section. As the Harvard Humanitarian Initiative flags, “data management comes with enormous legal and ethical responsibilities that most organizations are ill-equipped to handle, both in terms of systems and protocols, but also in terms of institutional culture and attitudes towards privacy.”50 Without proper data management protocols, humanitarian organizations, MNOs, or contracted third parties may be vulnerable to a number of problems affecting their CDR data. Actors could face a data breaches, which could result in the leaking of aggregate CDR or even personal data. CDR data could also be de-anonymized through triangulation with other data sources (the “mosaic effect”), leading to the identification of individuals or demographic groups.a, 51 These risks largely depend on what state of CDR data (raw, anonymized and aggregated, analyzed, etc.) is made available to the humanitarian organization, which is why we recommend a model where as much data is kept at the MNO as possible. Data retention is another potential issue, though this is also dependent on the state of CDR data a humanitarian organization is permitted to access. CDR may be useful for monitoring longer-term displacement or understanding general mobility trends to help with emergency preparation; however, this perceived utility could incentivize continuous data access and retention. According to the ICRC “Handbook on Data Protection in Humanitarian Action,” aggregate and anonymized data can be retained for a specific humanitarian purpose,52 but long-term data retention may increase the likelihood fa The Center for Humanitarian Data defines the mosaic effect as occurring when, “when multiple datasets are linked to reveal significant new information.”
Assessing the Use of CDR for Monitoring Mobility and Displacement