IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Government Data Needs an Ethical Foundation (Contributed)

While the private sector has struggled with unethical uses of data, government can learn from these mistakes and craft best practices by creating a data framework that is both principled and innovative.

In the last decade, we’ve seen amazing advances in technology that have made it easier for companies to create efficiencies, provide client services, and reach the ideal audience for their messages. With the rise of big tech, we’ve seen a surge in big data. This allows organizations unprecedented access to an individual’s personal information, with little control over how and why that data is used. Machine learning and predictive analytics are becoming more ingrained in our daily lives — the shows we watch, the social groups we join, and the transit routes we take are all based on unseen algorithms. 

While I’m excited for the new opportunities these technological advances present, I’m concerned we are losing sight of an important aspect of how we, as leaders in the public tech space, leverage new technologies while putting the customer first. We need to think about the ethics of data and what an ethical data program looks like. When creating an ethical data framework, we should fulfill these six principles: Ownership, Transparency, Openness, Privacy, Consent, and Literacy. 

While algorithms can have an arguably minimal impact on our daily lives, as the examples above suggest, they can also define someone’s future — including job application processes, higher education selection processes, and an increased use of facial recognition software in public spaces. While well-intentioned — and some not-so-well-intentioned — organizations believe that their data models are objective measures of reality, it’s impossible to escape the subjective. All of this comes down to the data we collect and how we use it. 

When thinking about government use of AI and advanced analytics, we have the responsibility to dictate how data is being collected, used and stored in order to mitigate potential impacts on our most vulnerable populations. And as governments increase their use of big data and algorithmic systems, we have a mandate to understand the unintended consequences this has on equity and social and economic opportunity for our residents.  

Technology is beginning to reach into even the dustiest corners of government back offices. We’re seeing a great variety of small, disparate systems and a handful of massive catch-all monoliths come online. These systems create a tangle of data. It’s a net positive that governments are modernizing and using tools to create efficiencies, but we need to think holistically around data management and ask ourselves, “we can do this, but should we?” 

So how do we do this? What guidelines, policies, and best practices can we put in place to shape the acceptable use of (often sensitive) data as we continue to evolve in this fast-paced tech world? 

Data Discovery as a Mindset

When standing up a customer relationship management (CRM) system, data system or program, the first step is to ask questions of your data and of your users. Often, the people setting up your systems are experts in the system, not in a particular use case. To make certain users are solving what they intend to solve, start by asking basic “who, what, when, where, why, and how” questions. This allows people to document the decision-making process and iterate where necessary in the future, while providing transparency.

Some examples include: 

  • Why are we collecting this data?
  • What information or data points are we collecting? 
  • How is this information being collected: Digitally? Handwritten and transferred? In an automated way?
  • Who is collecting it? Is an individual providing it? Is it being collected by a third party? 
  • How is it being used now? What are the potential future uses of this information? 
  • How is this information stored?
  • Who has access to the information? 
By implementing a discovery mindset, we can identify areas of risk, places where bias can be introduced or exacerbated and data points that may not be necessary to collect while providing a road map for future innovation and iteration. 

Governance as an Ethical Framework

The next step, which ensures the application of Ownership, Transparency, Openness, Privacy, Consent, and Literacy, is governance. By creating an intentional governance structure at the policy level, and providing every employee who touches data with clear, concise guidelines, we can start building strong foundations of not only data literacy but data ethics. 

An ethical data governance structure should incorporate the following principles:

  • Ownership – Who “owns” the data? Clarifying ownership, both at an agency level and a system level, provides a clear understanding of access and usage. 
  • Consent – Be explicit about what information is being collected and how it’s being used. Individuals should be allowed to easily opt out of data collection. 
  • Privacy – Ensure sensitive or protected information stays private. 
  • Openness – Share aggregate or row-level information with the public where appropriate (and keep in mind privacy and consent). Governance structures should account for publishing open data in machine-readable formats with dictionaries for ease of use. 
  • Transparency – Create transparency by not only documenting process and decision-making, but by documenting any code or algorithm used in conjunction with the data. 
  • Literacy – When talking about or displaying data, don’t use jargon or overly technical terms. Know your audience and work with them to create a common language around acceptable usage. 
(For additional reading, check out Andrej Zwitter and Neil Richards and Jonathan King’s work on Big Data Ethics.)

We’ve seen the detrimental impact of unethical uses of data and AI in private companies. Governments have to learn from the mistakes of others while not stunting innovation. We can take best practices, and lessons learned from others, and apply them to the public space. We know bias exists, we know seemingly benign decisions in data collection cause unforeseen consequences, and we don’t have to repeat the same missteps we’ve seen play out over the last decade. Government has a duty to be mindful and intentional in its data collection and usage practices. By creating an ethical framework within public institutions, we are poised for a future where government is seen as a trusted innovator when providing services for all residents, including the most vulnerable. 

Dessa Gypalo serves as the inaugural chief data officer for Cook County, Ill., where her focus is to bring a culture of data to the county while leveraging data assets to make informed decisions. Over the last decade, Dessa has worked with local and national political campaigns, and provided data support for numerous nonprofit organizations. She has led teams and initiatives at NGP VAN/ EveryAction, Enroll America, and Obama for America. Dessa holds a BA in History and Business Administration from Roanoke College.