Data ethics – establishing a solid code of conduct

Data ethics emerge as an axis of growing importance in the field of data science and research. No science can properly work without relying on the support and rectification provided by metaphysics. Data and Big Data, however, pose some specific challenges that need to be addressed and overcome.

How come something as abstract and seemingly neutral as data requires ethical considerations? The critical word in the latter sentence is “seemingly”. Data, and especially Big Data, is no simple collection of objective facts. It reaches beyond the borders of pure technology – if this can ever be conceptualised in such an encapsulated form – into the realm of human existence. In order to understand possible moral consequences of data research we have to move away from a familiar understanding of harm. The focus of data ethics lies on consent and privacy. The right to privacy is a human right and needs to be respected as such in any instance.

Consent is crucial in two respects. People have to be aware and approve of the collection and saving of their personal data in first place. What seems so obvious is not always that clear in reality, because only few people really read terms and conditions when agreeing to certain transactions. In many cases customers do not know about personal data being included in the transactions. Taking Facebook as an example, most users think of it as a cost-free application that is funded via advertisement, ignoring the fact that one actually allows the corporation to analyse and sell one’s private data in exchange.

The second level of consent requirement arises later when there are already sets of data in existence. The special characteristic of Big Data is its asset to combine and interpret data in such new ways that produce genuinely novel and unforeseeable insights. Therefore, a general conflict develops, as it is difficult to accept research practices whose outcomes are by no means curtailed. Also, collected data can thus be used for purposes the donor did originally not agree to.

Whereas consent and privacy are prime concerns, other possible repercussions of Big Data use must neither be overlooked. Poorly or improperly supervised analyses can exacerbate social and economic injustice and give greater weight to biases, since the possibility of identifying types paves the way for group discrimination. As a matter of fact, we need to realise that the subject of data is nothing neutral, but is made up by human beings. It is them giving away the initial information, feeding ongoing data systems and being confronted with the outcomes in their daily lives.

Due to this multilevel nature of data and its impact, also corresponding ethics have to span diverse layers. That is, certain universal principles have to be held up during the whole processing chain of data, which encompasses collection, aggregation, sharing, analysis, storage, monetisation and disposal. While some observers argue that it is too late for advocating moral integrity in the discipline of Big Data, many others stand up against this view and support the idea of a normative code of conduct guiding the way in this new and fast developing field. Taking the so-called Common Rule and Institutional Review Boards of universities as a model, a list of necessary principles can be established. It may contain the following aspects: [1]

• First of all, the actual people behind the data need to be respected at all times
• Second, the context and history of data sets need to be taken into account. Data must never be perceived as raw or pure.
• Third, privacy and security safeguards must meet the customers’ expectations. Proper information and clarification is key.
• Fourth, whoever conducts work with data, be it official institutions or private corporations, needs to comply with the law. Ideally, however, the law should serve as a minimum limit.
• Fifth, the highest possible degree of transparency shall be targeted. Methodologies have to be laid open and explained.
• Sixth, practitioners have to be accountable for their conduct, so no climate of a lawlessness can evolve.
• Seventh, practitioners always have to bear in mind the potential risks and dangers accompanying data. In order to counteract data’s tendency to work as a tool of exclusion and discrimination, this threat has to be recognised and actively tackled.
• Eighth, to take the former point one step further, not only shall no one be excluded, but those who create the data should be the ones benefitting from its use. People shall not become objects of data exploitation. Ownership is a prime keyword in this context.
• Ninth, it is reasonable to install external review institutions that prove research proposals before they are realised.
• And finally, tenth, it is sensible to abstain from collecting data without adequate reasoning and methodology, so to say just for its own sake.

Such a Decalogue of data can help current practitioners and lead future research in healthy ways. It is not naïve to belief in an actual implementation of such a normative turn. The authors of the book “Data Ethics: The New Competitive Advantage” [2] argue that respect for privacy and sovereignty will be a crucial feature in our modern market economy for businesses to survive. Just like corporations need to take ecology into account in order to keep up with their competitors and to attract clients nowadays, they will soon have to adhere to a consistent policy of data ethics when aiming at establishing themselves in the field.

References

1. Accenture technology (2017). Latest thinking. Building digital trust: The role of data ethics in the digital age. [online] Available at: https://www.accenture.com/us-en/insight-data-ethics [Accessed 8 Dec. 2017].
2. Hasselbalch, G. and Tarnberg, P. (2016). Data ethics: The new competitive advantage. Copenhagen: Publishare. [online] Available at: https://dataethics.eu/book/ [Accessed 8 Dec. 2017].

Keywords

data ethics, Big Data, privacy, consent, data sovereignty

Author: Andrea Ruscher