Case-Inclusive Transparency for a Digital and Open Government (CITaDOG)

Digitalization of government is not only potentially constraining public officials in taking tailormade decisions, it can also enable them to take decisions which are better apt for the particular circumstances at stake. However, this idea of administrative discretion in decision-making suffers from a wicked problem: the rule of law requires government decisions to be consistent with other decisions, but citizens have hardly access to cases other than their own.

Existing transparency guarantees, which merely focus on the publication of rules, make it difficult, if not impossible for citizens to compare their case with other relevant cases and therefore fail to address this need for so-called ‘case-inclusive transparency’. However, the current societal emphasis on open government constitutes a highly promising game changer by requiring enormous amounts of government documents, including single-case decisions, to be made public in a machine-readable way. Unfortunately, this potential is still highly unexploited due to a combination of legal, technical and behavioral constraints.

CITaDOG will assess to what extent public disclosure of single-case decisions as open data can strengthen transparency in administrative decision-making. This ‘case-inclusive transparency’ should empower citizens as watchdogs (‘citadogs’), thereby strengthening equality of arms between citizens and governments and increasing the legitimacy of administrative single-case decision-making.

Project objectives

CITaDOG distinguishes three key objectives as necessary building blocks for shaping the overarching concept of case-inclusive transparency.

• ‘Case Disclosure’: characterizing practices of public disclosure of single-case decisions under open government legislation;
• ‘Case Connection’: exploring to what extent patterns of decision-making can be inferred from single-case decisions disclosed under open government legislation;
• Case Impact’: examining whether and how public officials’ decisions differ after exposure to NLP-derived patterns of decision-making, made by other public officials.

CITaDOG will first (i) unveil how open government legislation steers current practices of public disclosure of single-case decisions. Next, using single-case decisions that are already available on Dutch governmental information portals, CITaDOG will (ii) develop and apply AI/NLP techniques with the aim of deriving patterns of decision-making from this enormous body of single-case decisions. Finally, CITaDOG will (iii) assess how public officials respond to this publicly disclosed information in their decision-making behavior by examining how they change their decisions after being exposed to different levels of information. Together, these three paths will further shape case-inclusive transparency and strengthen the informational and legal position of citizens vis-à-vis governments.

Potential impact

Digitalization of government is rapidly transforming administrative decision-making. The prevailing idea is that governments are increasingly taking resort to supportive or even prescriptive modes of automated decision-making, thereby gradually transforming into so-called ‘system-level bureaucracies’. The accompanying concern is that the outcomes of administrative decision-making in individual cases are increasingly determined by complex and non-transparent algorithms, which fail to take into account the particular characteristics of the case at stake. However, there is also an opposite perspective to digitalization of government, arguing that digital government could instead strengthen tailormade administrative decision-making, in particular through the use of open government.

By combining legal, technical and behavioral insights, CITaDOG will not only lead to a new understanding of transparency and its effects on decision-making behavior, which is indispensable for discretionary administrative decision-making. It will also develop the tools necessary for realizing meaningful transparency in an era of digitalization and guide existing attempts within explainable AI to ensure transparent and consistent decision-making. Finally, CITaDOG will generate concrete recommendations for governments on how to improve disclosure of information and reason-giving, while empowering citizens to access and analyze those decisions themselves.

The overall aim of CITaDOG to develop case-inclusive transparency can be considered a concrete manifestation of digital sovereignty of citizens in the context of administrative single-case decision-making. Case-inclusive transparency assists citizens to make sense of the enormous amount of publicly disclosed decisions and to strengthen their informational rights in decision-making processes.

Duration

The project will run from July 2023 to December 2027.

Multidisciplinary project consortium

As developing case-inclusive transparency is both a legal, technical and behavioral endeavor, these different perspectives are included in the research team that brings together early career and advanced scholars from three different schools: Tilburg Law School, Tilburg School of Humanities and Digital Sciences, and Tilburg School of Social and Behavioral Sciences.

CITaDOG consists of scholars from administrative law, public administration, social psychology and computational linguistics. Two PhD researchers will complement the research team, one having a socio-legal orientation, examining how the legal interpretation of transparency rights and the application thereof in administrative practices are affected by a more case-based perspective on single-case decision-making, the other having a research profile in legal analytics, developing new tools to apply NLP techniques to legal text which an often unstructured character.

The research of CITaDOG is further strengthened by its close cooperation with two other national initiatives. The first initiative WetSuite, led by Prof Dr Anne Meuwese (Leiden University), aims to provide a research infrastructure for legal scholars with a collection of tools for the application of NLP techniques to Dutch governmental legal data. The other initiative Woogle is a project led by Dr Maarten Marx (University of Amsterdam) which provides a central platform for the disclosure of government information and which cooperates with several municipalities.

Key stakeholders: citizens, governments and intermediaries

As CITaDOG deals with transparency in the relationship between governments and citizens across the rich variety of single-case decision-making (e.g. licensing, sanctioning, subsidizing), both ‘the government’ and ‘the citizen’ are considered key stakeholders in this project.

In line with these key stakeholders, the consortium has been composed such as to ensure the presence of these different perspectives in the research project, either as co-funder or as collaborating partner.

Project consortium

This project is funded by Tilburg University’s Digital Sciences for Society program: