The ethical dilemma of data sharing under risk

Fillia Makedon, C.B. Owen, C. Sudborough, S. Kapidakis, P. Gloor, C. Heckman, J. Ford, J. Pearlman


Cases where the failure to share information places humans or property at risk present an ethical dilemma: the risk of harm supports ethical arguments for less restrictive information sharing, while privacy and ownership concerns support ethical arguments for more restrictive information sharing. This paper considers technical means to address this ethical dilemma in the context of a new data-sharing framework called SCENS (Makedon et al. 2003, Ye et al. 2003) based on metadata libraries (Makedon et al. 2002).

The SCENS approach uses negotiation technologies to establish conditions for information sharing, allowing for sharing that would not be possible otherwise. The negotiation system enables communication and cooperation as it brokers agreements from sharing parties that determine who, what, how and for how long there is access to the shared information. Case studies from four domains are analyzed to examine similarities and differences. The central thesis of the paper is that SCENS can provide an appropriate means for easing, and possibly resolving, the ethical dilemma of data sharing under risk.

Ethical issues relative to information sharing often arise because available information can be used in an unplanned way or without the consent of the information owner. Indeed, significant legislation has been passed that codifies these ethical principles into law. However, considering only the viewpoint of data owners ignores the ethical issues introduced when a risk exists that data sharing could alleviate. In these cases, sharing of information may be needed to solve a problem involving the safety or well-being of humans or property.

In this paper we analyze cases in the context of four domains: medicine, social services, e-government, and international security. We argue that well-meaning but poorly-crafted laws that consider only the negative effects of information sharing, such as the Health Insurance Portability and Accountability Act of 1996 (HIPAA 1996) in the United States, often put individuals at risk. We consider risks such as terrorism that demonstrate the hazards of not facilitating the sharing of information across government agencies. We then outline a negotiation approach based on determining restrictions and stipulations on the use of data on a case-by-case basis.

Case studies in four domains

1. Medical

Recent HIPAA restrictions greatly hinder data sharing capabilities among authorized medical clinicians and researchers. For example, medical test results may be unavailable to a clinical specialist even when an emergency exists, such as a decision whether to operate on a patient or not. In this case, the desire to protect patient’s data from misuse may run counter to the needs of the patient.

2. Social services

Social services manage situations such as mental health, poverty, alcoholism, and child abuse. The data may include psychological observations, performance tests, or brain scans, and due to legitimate concerns about privacy and the sensitivity of individual records, different institutions rarely make these available except on a case-by-case basis. Thus, although sharing in this domain may dramatically improve the level and quality of care by cutting costs, helping research, and aiding funding efforts, it is currently very difficult to arrange due to the effort required for each attempt to gain authorization for access.

3. E-government

The new electronic communication infrastructure known as “e-government” has made many government services accessible through electronic interfaces, both for the public and for internal government affairs. E-government can offer around-the-clock access, transparency, accountability, up-to-date information, and other benefits. To be fully efficient in such a system, different government agencies must be willing to share their data and analyses of these data; however, concerns about data security and acknowledgement of effort often hamper cooperation.

4. International security

International security requires preparedness for current and emerging risks. Security organizations face difficult challenges overcoming cultural barriers and deriving knowledge from the assimilation of diverse information quickly and accurately. As the world becomes more crowded, complex, and interconnected, international security agencies operating in isolation are hampered in their efforts to monitor and assess security risks. However, as in the case of e-government, cooperation between parties is hampered when the benefits of making contributions are difficult to perceive.

SCENS sharing approach

We believe that a technological solution can be used to attack the ethical dilemma between satisfying concerns about security, ownership, and privacy and the perceived need to provide access to data needed to combat risks. Key to deciding on an ethical information sharing approach is to involve the people who provide and use the information in a negotiation on how it may be used. To address privacy concerns, people can consent on the use of personal information as part of the negotiated conditions: this can include determining what part of the information will be available for use in each case, in which cases information will be available, and who will be allowed to make use of the information (or what conditions the user of the information will satisfy). An information sharing system must include appropriate incentives for both information owners and information users.

A negotiation-based mechanism of communication and cooperation
At Dartmouth we are building SCENS: Secure Content Exchange Negotiation System. SCENS automates the process of reaching agreement between two or more parties as it documents the conditions under which exchange or sharing of information is to take place. These conditions can be usage requirements, such as limitations on duration, or specification of the persons or groups who can be involved in information sharing. SCENS also supports “group negotiation” useful for “distributed decision making”. Group negotiation sets conditions that a critical mass of users must agree upon, and allows individual negotiations to succeed only if many contracts with similar terms reach agreement.


U.S. Public Law 104-191. August 21, 1996. Health Insurance Portability and Accountability Act of 1996, subtitle F: Administrative Simplification.

Makedon F., Ford J. C., Shen L., Steinberg T., Saykin A. J., Wishart H. A., and Kapadakis S. 2002. “MetaDL: A Digital Library of Metadata for Sensitive or Complex Research Data,” presented at European Conference on Digital Libraries (ECDL2002), Rome, Italy.

Makedon F., Kapadakis S., Steinberg T., Ye S., and Shen L. 2003. “Data brokers: Building collections through automated negotiation,” Dartmouth College Computer Science Department, Hanover, NH, Technical Report DEVLAB-SCENS-03-02, March 2003.

Ye S., Makedon F., Steinberg T., Shen L., Ford J., Wang Y., Zhao Y., and Kapidakis S. 2003. “SCENS: A system for the mediated sharing of sensitive data,” presented at Third ACM/IEEE Joint Conference on Digital Libraries, Houston, TX.