In conversation with Olivier Thereaux: Exploring the field of data ethics
As part of our data ethics project, we sat down with Olivier Thereaux, head of Research and Development at the Open Data Institute. Jennifer and Susanne had a discussion with Olivier about ODI's purpose, the landscape of data ethics, and the challenges the field faces.
The beginning of the ODI's work
As Olivier mentions, it's important to first start with the 'why' and the 'how'. Data ethics has actually been around for quite some time. However, it was the onset of various Big Data scandals - such as Cambridge Analytica - in the last 5 years that truly catapulted data ethics into the spotlight. As such, we have generated an enormous volume of data ethics research and initiatives. At the forefront of this is the ODI's work - the Data Ethics Canvas and Consequence Scanning frameworks come to mind. Such tools are used all around the world with the purpose of ensuring that data actors are asking the right questions about stewardship, use and impacts.
But this was not an easy feat. Thinking about 'data ethics' and 'open data' had high barriers to entry, particularly in the early stages of discussion. One must consider the risks businesses face by opening up their data, discussing the risks of data practices, and by prioritizing data ethics when other issues (such as business operations) are deemed to be much more consequential.
One way in which the ODI tackled these questions is by designing data ethics toolkits based on already familiar frameworks. For example, the Data Ethics Canvas, which is a direct descendant of the widely-used Business Model Canvas and the Online Ethics Canvas, makes data ethics much more accessible for data organizations. Moreover, the decision to focus on data ethics rather than data justice was one that was carefully considered. Data ethics, which evaluates both positive and negative sides of data practices, proved much easier for people to get on board. As Olivier says, data ethics is not about perfection; rather, ethics is fluid and fast-paced, able to reflect the zeitgeist of what the organization thinks is right. In fact, Olivier stresses that businesses that take ethics beyond the base level of legislation could leverage this as a Unique Selling Point.
If we refer to the tech lifecycle, Olivier says that we are at the 'chasm' as coined by research by the Open Data Institute and Consequential - we are moving from the experimentation and early adoption phases to evaluation and polishing of existing tools. What have we learnt in the past decade and what challenges do we have to overcome?
One of the biggest challenges we talked about was inclusivity and participation. On this note, I thought it important to loop back to the concept of data justice, and in particular, structural and distributive justice: Making sure that data decisions are made with community involvement. Whilst meaningful engagement does require strong will and capabilities, past experiments show that the benefits are plentiful. For example, the Ada Lovelace Institute's recent public engagement project shows that many citizens are actually aware of data ethics concerns in regard to COVID-19 management, and are able to articulate thoughtful and sophisticated concerns. Another report by The ODI further finds that the general public actually understands and cares about issues much more than they are given credit for!
Another concern in the field is that certain ethics issues are overshadowing others: We must remember that not all data is personal. For example, data points such as carbon emissions are incredibly helpful in tackling the climate crisis and impacts the wider population. In general, to cross this 'chasm' we must show that data ethics don't necessarily require radical change or perfection; making ethical decisions can come in baby steps. Olivier proposes these slow and steady advancements as a way to avoid the dreaded "hype cycle" - a skyrocketing and just as quickly descending level of interest and involvement.
Hopefully, by data ethics gaining more attention and by data ethics toolkits becoming much more accessible, we can continue to grow in the areas of equity, engagement, and ethics. At ODILeeds we want to help make these data ethics baby steps by making existing work more accessible through our research and interviews, which you can find here. If you want to join our journey or have an idea for where we should go next, reach out to firstname.lastname@example.org.