What Comes After Harm? Mapping Reparative Actions in AI through Justice Frameworks

TL;DR

AI harms are increasingly scrutinized, yet reparative actions post-harm are understudied. The authors develop a justice-informed taxonomy of AI harm reparative actions and apply it to 1,060 incidents in the AIAAIC dataset, using purposive sampling and LLM-assisted coding. They find that most reparative actions are symbolic (acknowledgment and attribution) with few remedies or structural reforms, highlighting accountability gaps. The work advocates centering affected communities, expanding civil-society oversight, and grounding reparation in punitive, restorative, and transformative justice to achieve meaningful redress.

Abstract

As Artificial Intelligence (AI) systems are integrated into more aspects of society, they offer new capabilities but also cause a range of harms that are drawing increasing scrutiny. A large body of work in the Responsible AI community has focused on identifying and auditing these harms. However, much less is understood about what happens after harm occurs: what constitutes reparation, who initiates it, and how effective these reparations are. In this paper, we develop a taxonomy of AI harm reparation based on a thematic analysis of real-world incidents. The taxonomy organizes reparative actions into four overarching goals: acknowledging harm, attributing responsibility, providing remedies, and enabling systemic change. We apply this framework to a dataset of 1,060 AI-related incidents, analyzing the prevalence of each action and the distribution of stakeholder involvement. Our findings show that reparation efforts are concentrated in early, symbolic stages, with limited actions toward accountability or structural reform. Drawing on theories of justice, we argue that existing responses fall short of delivering meaningful redress. This work contributes a foundation for advancing more accountable and reparative approaches to Responsible AI.

Figures (1)

• Figure 1: Distribution of Reparative Actions by Justice-Oriented Goals. Perpetrators' communication was most common, while legal or policy change, compensation, and product discontinuation were among the least frequent, each appearing in fewer than 8% of cases.