Rights-Based Approach to Artificial Intelligence Design: Addressing Harm through Participatory ex ante Impact Assessment

Authors: Vanja Skoric

Abstract:

The paper examines whether the impacts of artificial intelligence (AI) can be meaningfully addressed through the rights-based approach to AI design, investigating in particular how the inclusive, participatory process of assessing the AI impact would make this viable. There is a significant gap between envisioning rights-based AI systems and their practical application. Plausibly, internalizing human rights approach within AI design process might be achieved through identifying and assessing implications of AI features human rights, especially considering the case of vulnerable individuals and communities. However, there is no clarity or consensus on how such an instrument should be operationalised to usefully identify the impact, mitigate harms and meaningfully ensure relevant stakeholders’ participation. In practice, ensuring the meaningful inclusion of those individuals, groups, or entire communities who are affected by the use of the AI system is a prerequisite for a process seeking to assess human rights impacts and risks. Engagement in the entire process of the impact assessment should enable those affected and interested to access information and better understand the technology, product, or service and resulting impacts, but also to learn about their rights and the respective obligations and responsibilities of developers and deployers to protect and/or respect these rights. This paper will provide an overview of the study and practice of the participatory design process for AI, including inclusive impact assessment, its main elements, propose a framework, and discuss the lessons learned from the existing theory. In addition, it will explore pathways for enhancing and promoting individual and group rights through such engagement by discussing when, how, and whom to include, at which stage of the process, and what are the pre-requisites for meaningful and engaging. The overall aim is to ensure using the technology that works for the benefit of society, individuals, and particular (historically marginalised) groups.

Keywords: rights-based design, AI impact assessment, inclusion, harm mitigation