Data ethics, it's cricket

For an issue that has become so topical, our prevailing data ethics conversations surprisingly say little about the ethical principles that underpin them. There’s much said about the importance of transparency, accountability and harm minimisation, but little about why and how these concepts actually contribute to ethical behaviours and outcomes.

Take transparency. It is, in fact, ethically agnostic. It is “proto-ethical”, a precursor that enables ethical practices, but not necessary.

Take accountability. The notion of accountability is important, but without a means of giving expression to the term, it is a “scape” waiting for a “goat”.

And harm minimisation. In big, linked, integrated and shared data settings that house semi-autonomous and opaque algorithms, how do you determine the direct and consequential potential for harm? Not just the immediate risks, but unforeseeable future ones that may be caused by data drift and changes in algorithmic behaviour?

So here’s the cricket bit. Well at least in a normative ethics kind of way...

Let’s imagine I’m playing cricket and snick the ball, which is then caught and appealed by the wicketkeeper. Do I walk or wait for the umpire’s finger to go up?

  • The ethics of virtue and character might suggest that I walk as an expression of my “cricketing values”.
  • The ethics of duty and rules might suggest that I stand my ground and wait for the umpire to give me out, or not.
  • The ethics of consequences and results might suggest that by walking I may be letting the team down, but also should I be given not out, that it might reflect poorly on my character.

If my objective is to win at all costs, how far might I be willing to go to achieve it? I might resort to sledging opposition players, which is not against the rules but is of questionable virtue. I might tamper with the ball, which is against the rules but may provide the winning edge. Do the ends justify the means? If I narrow my consideration of the consequences of the need to win, the risk of getting caught might seem worth taking. However, if I were to consider the wider consequences of reputation damage and penalties, I may consider the risk too great.

From an even wider perspective, if cricket administrators considered the consequences of fostering a “win at all costs” culture, might they have reasonably been able to foresee the potential for a ball-tampering incident to arise?

The ethics of data and algorithms decision-making present us with a far more complex and ambiguous landscape to navigate. Notions like transparency, accountability and harm minimisation can only meaningfully exist within a wider ethical context. Without them, they are merely set of principles in search of expression.

So, if you’re accountable for data governance or responsible for its management, here are three things you need to be able to demonstrate:

  1. That your data operations serve a purpose that is legitimate and consistent with your organisation’s stated values and principles.
  2. That you have considered the consequences to the extent that might reasonably be expected.
  3. That you have complied with all relevant legal and regulatory requirements.

After all, it’s cricket.

DISCLAIMER: We want to foster discussion on data issues that are relevant to all Australians, by collaborating with experts in the private, civil society, research and government sectors to publish their posts and linking to their articles on our website. The views expressed in the posts and comments are not necessarily those of the Office of the National Data Commissioner. We encourage you to add your views in the comments section of the post.

Contribution Privacy

This article has 0 comments