When I tell people that I work on data governance, the next question is usually ‘What’s that’?’. While data governance shapes the world of tech around us, it isn’t necessarily intuitive as a concept.
Here, I’ll show what a broad definition of data governance is, how it is useful, and why talking about ethics is not enough. For practitioners working in tech, a broad definition of governance can build a shared understanding to align projects with the public interest.
A broad definition of governance
In conversations around public interest tech, there is a lot of talk about the ‘why’; which can be normative mission statements like ‘don’t be evil’, or more abstract values, like transparency. It’s also common to focus on the ‘what’; the details of a project or a particular piece of technology. It’s less obvious to discuss the nitty gritty details of the ‘how’. This is where governance comes in, connecting the various disconnected parts.
Governance includes laws, ethics, procedures, how these interact, and how we talk about these things.
This is different than narrower definitions in different disciplines. Governance sometimes means working together with a range of actors outside of government. Sometimes it means the politics of institutions. Sometimes it is about the operational procedures within a corporation.
As we wrote in our paper:
"A broad definition of data governance is not only a set of requirements, procedures, architectures, or laws. Governance is the sum of these parts, as well as the discussions and negotiations to get there, the process of working together and understanding what the things we have created actually do, and whether those functions achieve the things we want them to achieve. "
I work on data governance, though the broad definition of governance can apply to the digital or to tech as well, depending on what the object is you’re looking at.
This entire normative map draws from the work of Luciano Floridi, an information philosopher. His open access paper soft ethics and governance of the digital is particularly useful.
From the perspective of social science, governance also includes the way we communicate shared understandings through common languages, jargon, discourse. The perspectives we use determine what we’re able to see and whether we have an understanding we can act on.
In practice, ethics is not enough
Mentioning ‘tech ethics’ can kick-start a conversation, because it gets us thinking about norms and values in relation to tech. The challenge is, ethics means too many different things to different people, and that vagueness is unhelpful. While the term is a marker for norms and values, tech ethics today doesn’t necessarily build a shared understanding.
Ethics as a political philosophical discipline has a long and rich deliberative tradition. In practice, however, my colleagues and I found that ‘tech ethics’ was being used in a much narrower way.
In our recent research on governing public interest technology in cities, we found three particular ways that ‘ethics’ functioned in conversation.
First, ethics as a checklist. Where you tick off the boxes and that you are deemed ethical and have fulfilled your auditing requirements.
Second, as a bureaucratic behemoth; a process so huge and so slow that it evolved to nightmarish proportions, which required approval from many different people on many different echelons of the hierarchical ladder. In this case, this was the only process for formal ethical approval the organisation had, and yet the process was so cumbersome that nobody wanted to do it and would much rather it rested quietly in a drawer.
Lastly, we encountered ethics as a sort of ethics-washing and marketing, where everybody involved knew it was about getting the rubber stamp of approval so a project could continue regardless, rather than a genuine engagement with the questions around the public interest. As independent researchers, we were adamant to distance ourselves from these kinds of expectations.
When ethics is used in these narrow ways, it becomes empty of people. There are no humans who have perspective; neither in the landscape, neither within tech projects.
Expanding our vocabulary
In the context of tech in the city, ethics needs to be complemented by practical considerations of what happens and how it happens, rather than abstract values or technical specifications alone.
The challenge is, discussions often don’t have the language that can handle a wider vision. We simply don’t know how else to describe the things we value other than ‘ethical’.
As a result, the conversation and its ensuing solutions get limited to a narrow understanding of ethics.
Governance creates an understanding of what is actually happening.
To be more effective in designing systems and practices to guard the public interest, we need to expand our vocabulary. By finding ways to describe targets and goals that go beyond ethical checklists and legal compliance, we can open up new possibilities.
New possibilities become visible
With this broad definition of data governance, it becomes possible to see how things interact and what the gaps are. As a result, it is possible to see a wider range of issues and risks that a particular piece of technology raises.
For example, in our work with city municipalities, we found that there tended to be an overwhelming focus on legal data protection compliance, and fuzzy — or non-existent — processes for defining and protecting the public interest.
There was a clearly articulated focus on the inputs to governance arrangements, but few mechanisms for monitoring and adapting to changes in technology use over time.
Remedies to these sorts of issues can then be designed based on the local context. As but one example, existing accountability mechanisms can be repurposed or realigned.
The process of identifying what issues can be seen and addressed with a broad definition of data governance is what we do in the Data Governance Clinics. Learn more here.