Towards Transparency by Design for Artificial Intelligence

Abstract

In this article, we develop the concept of Transparency by Design that serves as practical guidance in helping promote the beneficial functions of transparency while mitigating its challenges in automated-decision making (ADM) environments. With the rise of artificial intelligence (AI) and the ability of AI systems to make automated and self-learned decisions, a call for transparency of how such systems reach decisions has echoed within academic and policy circles. The term transparency, however, relates to multiple concepts, fulfills many functions, and holds different promises that struggle to be realized in concrete applications. Indeed, the complexity of transparency for ADM shows tension between transparency as a normative ideal and its translation to practical application. To address this tension, we first conduct a review of transparency, analyzing its challenges and limitations concerning automated decision-making practices. We then look at the lessons learned from the development of Privacy by Design, as a basis for developing the Transparency by Design principles. Finally, we propose a set of nine principles to cover relevant contextual, technical, informational, and stakeholder-sensitive considerations. Transparency by Design is a model that helps organizations design transparent AI systems, by integrating these principles in a step-by-step manner and as an ex-ante value, not as an afterthought.

Ethics-Based Auditing of Automated Decision-Making Systems: Nature, Scope, and Limitations

“…Throughout this purpose-oriented process, various tools (such as software programs and standardised reporting formats) and methods (like stakeholder consultation or adversarial testing) are employed to verify claims and create traceable documentation. This documentation process enables the identification of the reasons why an ADMS was erroneous, which, in turn, could help anticipate undesired consequences for certain stakeholders and prevent future mistakes (Felzmann et al, 2020). Naturally, different EBA procedures employ different tools and contain different steps.…”

“… 1 Formal (as opposed to informal) governance mechanisms are officially stated, documented, and communicated by the organisation that employs them (Falkenberg & Herremans 1995 ). …”

Technological Answerability and the Severance Problem: Staying Connected by Demanding Answers

“…This is sometimes framed in terms of ex ante explanation, or transparency by design (e.g. Felzmann et al, 2020 ; Rossi & Lenzini, 2020 ). For example, we might design self-driving cars so as to maximize fuel-efficiency above all other factors.…”

AI Ethics: Algorithmic Determinism or Self-Determination? The GPDR Approach

“…The ability of AI systems to make automated and selflearned decisions enlightens the need for transparency in the way that such systems reach decisions [52]. Machine learning algorithms have changed the way patterns are extracted from datasets and how predictions and decisions are made [53].…”