DATA: VALUABLE BUT NEVER NEUTRAL
By Ania Karzek, Manager Strategy & Governance, City of Holdfast Bay
“Data is the new oil” is a phrase attributed to mathematician Clive Humby. It’s a powerful metaphor, not only because it underscores the immense value of data as a resource, but also because it reminds us that, like crude oil, data must be processed and refined to become truly useful.
We often focus on the opportunities that data enables: evidence-based decision-making, personalisation of services, real-time monitoring, joined-up systems, predictive analytics, and more. The possibilities appear endless, limited only by our imaginations. Data promises to help us work smarter, act faster, and serve better.
But how often do we pause to consider the shadow side of data?
Beyond the surface similarities of oil and data, they share deeper characteristics. Both:
- Come in varying levels of quality
- Can be monetised
- Require specialist infrastructure
- Drive whole industries
- Despite being widely used, are concentrated in the hands of the few
- Can result in harmful consequences when uncontrolled or misapplied, and
- Can be exploited and weaponised.
Most critically, data is never neutral.
Every bit of data is a product of its context and our decisions – how to define it, what to measure, who gets included or excluded, and those decisions are shaped by human values. Our values, and those of our technology vendors, may not be explicit, visible, aligned or shared.
While a single piece of data may seem harmless, in combined data sets and the systems built upon them, its power grows exponentially. We do not gather data randomly – what we collect is framed by our goals, and when those goals are flawed, short-sighted, or politically charged, the consequences can be devastating.
Robo-debt offers a sobering example; what began as an attempt to save government money through automation ended in widespread harm, due to flawed assumptions embedded in the data and the systems interpreting it. Even when intentions are good, outcomes can be disastrous.
Bias is inescapable. What gets measured gets treasured, and what doesn’t get measured gets overlooked. For example, street-level crime is widely tracked and reported, but white-collar crime is often ignored. We cling to GDP as a measure of success while sidelining environmental or social wellbeing. Even “accurate” data can reinforce injustice - predictive policing models based on historical crime data being a prime example.
Most importantly, data both fuels power and shapes it. How data is framed, interpreted, and used can shape public opinion, justify policy decisions, or entrench systemic bias. The 2007 Northern Territory Emergency Response, for instance, relied on selective interpretation of data to justify sweeping, controversial actions. Government overreach is a real danger, easily fuelled by narrative-driven data; the fact that most of us have not had to give it much thought speaks to our privilege.
Data offers enormous potential, but only if we respect its complexity and treat it with the care citizens should demand. It is not a neutral input; it is a socio-technical construct shaped by context, intention, and power. As data continues to drive public policy, digital services, and AI systems, we must move beyond blind trust in its objectivity. Instead, we must ask ourselves questions such as: why are we collecting the data we’re collecting, do we have explicit or social license to do so, have we been transparent about how we’ll use it, who decides what matters and what doesn’t, who benefits or is harmed by how it is used, and how will we prevent unintended consequences from the data (ie, power) we’re amassing? The answers to these questions (and more!) should guide not only how we build data-driven systems, but how we govern them strategically, safely, ethically, and with humility.
Have you seen bias influence data? Can you think of an instance when your bias has impacted the data you've been working with? Start or join a discussion on the BPIN Viva Engage site to help build local government sector capability.