Insurers may have data provenance as part of their data ethics programme, but have they scoped it correctly? An often overlooked aspect of data ethics has recently received media attention. It deals with an age old problem in supply chains – how labour is managed and remunerated.
Hold on, many of you will be saying. We’re talking about digital here. Where’s the labour in data? Well, it’s certainly doesn’t occupy a front of house position, but then, that’s often why supply chain risk is difficult to tie down. It’s invariable ‘behind the scenes’.
Let’s start by clarifying what we mean by data provenance. It’s the documentation showing where a piece of data comes from, the processes and methodology by which it was produced and how it has been altered or transformed during its lifetime. So for example, wearing my ethics hat, your data provenance should be able to tell me what form of consent attaches to any of my data you use, plus when and who I gave that consent to.
The problem is that most people have a tendency to focus on three aspects of data provenance: the data, the technological processes and legal stuff like consent. However, lying behind all this are people. And that is where the supply chain risk comes in.
A lot of that original processing can involve people collecting data, interpreting it (for example, an image), labelling it and categorising it. And probably a lot more than this, but these are the obvious ones. Such work is needed because not all data is neat, clean and oven ready.
And this is often much more than a bit of tweaking here and there. The overall process of cleaning and preparing data is a huge part of a digital operation. Not all of that will involve people, but a large chunk of it will.
What the media is picking up on now is what they are calling microwork. This is internet based labour involving tasks that can last anything from a few seconds to a few minutes. Many such tasks involve some form of data handling.
The pandemic has drawn a lot more people into microwork and the cost of living crisis will add to this. The main issue involved is that pay is often well below the minimum wage and sometimes, in as many of one in seven cases, even unpaid.
Now, some of you will be jumping up and down, pointing out that under UK legislation, piece work by people not classified as employees doesn’t qualify for the minimum wage. And that’s right, but then that’s the legal perspective. The sentiment perspective is different.
We’re now seeing UK based investors campaigning for retailers to pay the living wage not just to their employees but to contractors as well. And these are not just niche investors, but big players like HSBC, Legal and General and Fidelity. It may not be long then before investors turn their attention to the minimum wage for people relying on microwork. It's already something that Uber has had to deal with.
The key point to remember here is that firms don’t want to be associated with exploitative labour practices in their supply chains, whether the work is in the UK or abroad. And it’s not difficult to see paying less than half the minimum wage as exploitative.
Questions for Learning Curves
So the questions for insurers are these:
- Exposure - do you know what it is and where it lies?
- Expectations - have you set them and are you communicating them?
- Mitigation – how is this being organised and do you know if it is working?
There’s little to no rocket science in any of this. What it rests on however is seeing the problem and taking ownership of it.
Twenty years ago, clothing retailers struggled to see the problem and take ownership for exploitative labour practices in their garment supply chain. Now everyone on the high street takes it seriously. Like it or not, digitally transforming firms are on a similar learning curve for their data supply chains.