Avoiding Pitfalls when Developing Data Ethics Principles

Ethical principles for data and analytics can be sourced in two ways. One is to take your firm’s existing ethical values and interpret them in a digital context. Another is to develop some stand-alone and more specific principles dedicated to data ethics issues. In both cases, there is one big issue to look out for at this initial stage.

The big issue to look out for is about using principles that have nothing to do with ethics and labelling them as ethical principles. So for example, you might have a corporate value like innovative and include it as an ethical principle, on the basis that innovation is a good thing for digital firms to do. The problem is that innovation is a business value, not an ethical value.

Who cares, some of you may ask. Well, your senior management should care, on the basis that if they’ve investing in building a credible position on the ethics of their digital transformation, building that credible position around non-ethical values is a waste of time. A bit like loading fuel onto your boat when its only got a sail. Both signal ‘they may not have done this before’ and a waste of effort and resources.

A Model for Ethics and Trust?

The big four consultancy Deloitte recently launched its first ‘State of Ethics and Trust in Technology’ report and in it, they set out seven ethical principles, as follows:

Safe and Secure - The technology is protected from risks that may cause individual and / or collective physical, emotional, environmental, and/ or digital harm.

Private - User privacy is respected, and data is not used or stored beyond its intended and stated use and duration; users are able to opt-in / out of sharing their data.

Responsible - The technology is created and operated in a socially responsible manner.

Robust and Reliable – The technology produces consistent and accurate outputs, withstands errors, and recovers quickly from unforeseen disruptions and misuse.

Transparent and Explainable – Users understand how technology is being leveraged, particularly in making decisions; these decisions are easy to understand, auditable, and open to inspection.

Accountable – Policies are in place to determine who is responsible for the decisions made or derived with the use of technology.

Fair and Impartial - The technology is designed and operated inclusively in an aim for equitable application, access, and outcomes.

How Ethical are They?

So how much ethics is there in each of these seven ‘ethical principles’? Well, in my opinion, not a lot and certainly not enough. Let’s look at each in turn. My reference to ‘score’ relates to how ‘ethical’ I found each principle.

Safe and Secure – this appears to be about ethics because of various harms referenced, but strangely, these harms are orientated around the technology being protected from them, not the consumers who engage with that technology. Score – difficult to judge, so poor.

Private – as everything in this principle is what UK and EU data protection legislation expect companies to do, the ethical ‘added value’ is minimal. Score – poor

Responsible – this principle references themes like humane, the common or social good, sustainability and added value, which can all be associated with ethics but are somewhat vague for a specific principle. Score - middling

Robust and Reliable – this is all about business process. Score – zero

Transparent and Explainable – the themes within this principle are all either in existing law or in the process of being incorporated into new law. Some like ‘easy to understand’ and ‘open to inspection’ are usually seen as unattainable, so this principle looks nice but lacks conviction. Score – middling

Accountable – this is all about business process. Score – zero

Fair and Impartial – this is worded quite differently to the others, in that it is the only one positioned as ‘an aim’, as opposed to ‘an actual’ which the others are. So it looks good, but rests on weak foundations. Score – I’ll be a generous and say middling

Why do Firms Struggle?

These are of course my subjective views, but even so, an overall score of three at middling, two at poor and two at zero is not great. None of these principles came across as firmly ethical. That doesn’t mean that nothing ethical will come out of their use, but as a starting point, they come across as more about business than about ethics, as somewhat of a wasted opportunity.

So why is this so? Why do firms often struggle with this? I believe there are two reasons.

The first reason is that firms often look in the wrong place for their ethical principles. They look at external hard lines like mandatory guidelines and regulatory frameworks. Then they look at slightly softer things like professional and sectoral standards of conduct. Then there are cultural attributes within their own and like-minded firms, about how they like to do things.

I’m not saying that these have nothing to add to the development of a set of ethical principles – they tell the firm things about the wider ‘data and trust’ landscape. The problem is that none of them tell the firm anything directly about what consumers would like to be happening. This matters because what is the point of a set of ethical principles if they aren’t built around the people for whom your firm is doing data ethics in the first place?

This is an absolutely key point to remember. Who are you doing data ethics for?

Outwards Looking

The second reason why ethical principles around data and analytics often seem to miss the mark or lack substance, is that they’re driven by the interests and priorities of senior executives. To a degree this is understandable – they’re the ones paying for all this, so why shouldn’t they have their say?

The problem is that your firm should not be doing data ethics for its senior executives. It should be doing it for its customers, prospects and partners. Ethics must be an outwards looking initiative, not an inwards looking one, if it wants to achieve the impact and influence expected of it.

Having your firm’s senior executives deliver the right ‘tone from the top’ for its data ethics initiative is of course very important, but that should be all about resourcing the initiative’s design and communicating its importance.

Short Shelf Life

Now some of you will say that this is only natural, that business people know more about business things and less about what consumers think. They’re more comfortable that way. Well, that’s true to a degree, but only a small degree. Businesses that don’t know their customers, that don’t listen to them, are businesses that tend to have a short shelf life. And the same holds for ethical principles built along the same lines.

It can be a strange, sometimes disconcerting experience for firms to reach out and be guided on ethics by their customers. Yet to coin some phrases - if not them, who? If not now, when?

In my experience, when I worked full time in a purely insurance role within the market, it can become very empowering to work that way. I’ve seen it cause organisations to do things no one thought they would do, that they never thought they would do. And to do those things with a clear belief that that was what they should do.

Two Questions

To sum up then, firms looking to organise a data ethics initiative around a set of ethical principles should always keep these two questions in mind:

• Is this actually an ethical principle or is it really a business one?
• Who is this principle designed to satisfy?