|
Article Excerpt
Abstract. This paper addresses the question of delegation of morality to a machine, through a consideration of whether or not non-humans can be considered to be moral. The aspect of morality under consideration here is protection of privacy. The topic is introduced through two cases where there was a failure in sharing and retaining personal data protected by UK data protection law, with tragic consequences. In some sense this can be regarded as a failure in the process of delegating morality to a computer database. In the UK, the issues that these cases raise have resulted in legislation designed to protect children which allows for the creation of a huge database for children. Paradoxically, we have the situation where we failed to use digital data in enforcing the law to protect children, yet we may now rely heavily on digital technologies to care for children. I draw on the work of Floridi, Sanders, Collins, Kusch, Latour and Akrich, a spectrum of work stretching from philosophy to sociology of technology and the "seamless web" or "actor--network" approach to studies of technology. Intentionality is considered, but not deemed necessary for meaningful moral behaviour. Floridi's and Sanders' concept of "distributed morality" accords with the network of agency characterized by actor--network approaches. The paper concludes that enfranchizing non-humans, in the shape of computer databases of personal data, as moral agents is not necessarily problematic but a balance of delegation of morality must be made between human and non-human actors.
Key words: actor--network theory, artificial agents, data protection law, delegation, distributed morality, intentionality, privacy
Introduction
Can morality, and in particular, privacy, be delegated to a machine or must it remain a purely human attribute? If we decide that we can delegate it to a machine, can we distribute aspects of morality through the network of machine and human in a balanced way?
In particular, I look at an example of delegation of morality where privacy is the central moral concern; the question of delegating morality to a database and how this continues to be achieved, successfully or otherwise, through the application of data protection law. This is considered in a past tragic case (Soham murders). One way of analysing key aspects of the case involves invoking a failure in the delegation of morality to a machine, where an appropriate approach to privacy was not achieved. However I suggest that some responses to this case (and the similarly tragic Climbie case), in terms of calls for the construction of a national database for children in the UK, can be read in terms of further attempts to delegate morality to a machine, despite the failures signalled by these tragic cases. I argue that we need to think carefully about the conditions where we would want to perform such a delegation and their implications. This is clearly complex and the examples to which I allude, suggest that we do not entirely understand the implications of the delegations we currently perform. The paper opens with a description of the Soham and Climbie cases, set against a consideration of UK data protection legislation, and the responses of the UK government in terms of its legislation to construct a national database for children. The concerns that prompt the calls for such a database are echoed in popular responses to protect children using microchip tracking devices. This is followed by a consideration of how moral agency, particularly concerning privacy could be part of a database, as opposed to treating the design and implementation of such a database in a purely instrumental way. This opens up a consideration of how delegations of morality might be treated theoretically. In formulating an analysis of such delegations three approaches are relevant; those of Floridi and Sanders, Collins and Kusch, Akrich and Latour. (1) The question of intentionality, which has beset philosophical critiques of artificial intelligence when delegation of intelligence is considered, is discussed but not found to be relevant. Instead, the concept of distributed morality, articulated by Floridi and Sanders but lent support from Akrich and Latour appears to be an appropriate approach for analysing the moral role of information and communications technologies.
The UK Data Protection Act--Soham and Climbie cases
The Data Protection Act (DPA) became law in the UK in 1984 and was subsequently superseded by the 1998 DPA which broadened the scope of the original act from just digital media to include data held by any means of storage including paper. That it was revised relatively quickly from its original form is an indication of how quickly our ethical and legal judgments must move in response to the use of computer technologies and how much our understanding of key concepts such as privacy are shaped by our use of new technologies. The DPA's application to personal data and its subsequent amendment to apply to non-computerized data serves to underline the way that new technologies, with their moral implications, quickly become entwined in our social existence.
The UK DPA applies to personal data and this recognizes that individual data privacy is important and must be protected and that individuals have a right to know what data is held on them, with the concomitant right to have the data changed if it is in error. Importantly, personal data should not be held longer than is necessary.
Two tragic, high profile cases which have arisen in the UK in the last ten years illustrate what can go wrong and where our moral intuition that the need to protect children is paramount is not reflected in the application of data protection law in protecting personal data. The common features of these cases relate to the way that they involve vulnerable members of society, in this case young children, whose need for protection and just treatment is uncontroversial. Secondly, both cases involve a failure to share personal data, about certain of the individuals involved, with other appropriate agents who might have ensured their protection, signalling, to a greater or lesser extent a failure to understand and apply data protection law appropriately.
Although, in general, personal data about individuals must not be shared with other organizations, there are exemptions for potential criminal activity, terrorism and the protection of vulnerable people. This is an important extension of the concept of privacy, one that is intended to be captured by data protection legislation, but which was not implemented appropriately in these cases. A major part of the concept of privacy involves informational privacy or the right to have information about oneself remain confidential. (2) Much, but not all, of the scope of data protection involves the protection of informational privacy. However, it is clear that there will be times when protecting the informational privacy of one individual might lead to a violation of decisional privacy i.e. the freedom to make and act upon one's own decisions without interference, or even physical privacy or freedom of unwanted physical observation or bodily contact or invasion. The UK DPA allows for normal informational privacy of an individual to be abrogated if a crime may be committed. In these cases, the crime involved was the worst example of violation of physical privacy, namely murder.
Victoria Climbie was sent...
|
