19 October 2018

The urgency for data ethics in public administration

Bron: Utrecht Data School
Source: Utrecht Data School

The Utrecht Data School (UDS) was present at the annual congress Overheid 360°, a platform for data-driven solutions in public administration. UDS was invited to discuss data ethics.

Iris Muis and Dr Mirko Tobias Schäfer gave a talk on data ethics and government, where they emphasized how data projects can constitute ethical dilemmas or plainly damage ones reputation or worse, damage citizens.

Bron: Utrecht Data School
Source: Utrecht Data School

myths of anonymized data sets

In their talk they debunked myths of anonymized data sets, demonstrated how data visualization steers interpretation, and discussed how bias in models and data affect the working of algorithms. They invited their audience to engage in a thought experiment, a not so fictional case of a Dutch municipality replacing their old streetlights with smart streetlight. What sounds like a great smart city initiative with enormous potential, also raises issues concerning private ownership of a public infrastructure, questions concerning ownership and access to data collected by the smart streetlights. A member of the audience shared that her municipality indeed had purchased some smart streetlights. "The deal was just too tempting, but later we noticed that we have neither access to the data nor do we own them.”

Data Ethics Decision Aid

The urgency for ethics as a means of navigating through the many issues and the complex long term effects of data projects became clear in this and other responses from the audience. On their stand, Utrecht Data School staff informed the many visitors about the Data Ethics Decision Aid and how it can be of use when developing data projects.

Ethical issues

The Data Ethics Decision Aid is a process for structurally assessing data projects in view of ethical issues. It helps development teams, project managers and decision makers to detect and address possible risks and ethical pitfalls. Developed in close cooperation with the City of Utrecht and their data analysts, DEDA functions as a dialogical tool for brainstorming, reviewing and deliberating data projects.

Consortium on data ethics

Licensed to the Association of Dutch Municipalities, it has been used for data ethics trainings at more than 50 municipalities. But there are also other areas where DEDA has been used (health and journalism). This year, UDS started to work with corporations and industry associations to test DEDA for implementation. An IOS Seed money grant supports to develop this initiative for building a consortium on data ethics.

DEDA as a research tool

While DEDA is a great tool for reviewing data projects, it also doubles as a research tool for participatory observation. UDS researchers utilize their DEDA workshops to learn more about how organisations work with data, which data projects they develop, and how they deal with data ethics. Recently, UDS received an IOS seed money grant for developing a research programme revolving around DEDA as method for analysing how data practices transform organisations. Currently, they work together with the Ministry of General Affairs in developing responsible data practices.

Bron: Utrecht Data School
Source: Utrecht Data School

Overheid 360°

Attendees at the congress Overheid 360° were information advisers, IT experts, process managers, data analysts from Dutch municipalities, the tax revenue service and the Netherlands defence forces. In general experts and decision makers for developing and implementing smart city solutions, data-driven management and innovation at government organisations. This event has grown significantly over the past two years. More than 30 companies and organisations presented their products and services in the business fair and the conference programme consisted of over 60 talks and presentations.

datafied society

Visiting professional congresses and business fairs is important for the methodological approach at Utrecht Data School. There they can meet partners for new research projects and they can effectively disseminate research findings to the group that is actually shaping the datafied society.