Measuring AI harms with multidimensional Lorenz Zonoids

arXiv:2609.16004v1 Announce Type: new Abstract: While AI systems increasingly shape high-stakes societal domains, their governance is limited by the lack of risk management methods that operate on real harms, taking their severity, and not only their likelihood, into account. As a consequence, AI risk management models remain compliance-driven and provider-centric, offering limited insight into how harms are dangerous, and on what should be the priority of intervention. The problem is amplified by the nature of harm data which are typically ordinal and multidimensional. To solve the problem, and offer an effective risk assessment methodology, in this paper we propose to model harm data by means of Lorenz Zonoids and Gini indices. To this aim we propose to extend them in a multidimensional setting, and show how to practically calculate them for a real AI incident data repository, provided by the Massachusetts Institute of Technology. The empirical findings indicate that environmental, infrastructure, property, physical, and democracy-related harms attain the highest values under the two multidimensional Gini indices and therefore exhibit the strongest concentration in their joint direct, indirect, and inferred severity-frequency distributions. These concentration patterns may help identify categories that warrant closer examination when mitigation priorities are determined.

Proposes new methodology for AI risk management.

The article proposes a new method to assess AI harms more effectively by considering severity and multiple dimensions of harm. This can improve risk management and prioritization but does not in itself guarantee reduction of harms. It is a technical advancement with mixed societal implications depending on uptake and enforcement.

This is a methodological contribution to AI risk assessment, not a claim about societal outcomes itself. It implies incremental improvement in governance tools by highlighting which harm categories are most severe, a modest positive for oversight capacity.

The paper identifies significant concentrations of harm severity across various critical societal domains due to AI incidents. The findings highlight existing, real-world harms rather than potential future ones, providing a quantitative basis for prioritizing interventions against these risks. This directly implies serious negative consequences for human flourishing.

The paper highlights significant harms from AI systems in critical areas. The current risk management models are inadequate, focusing on compliance rather than real harms. The proposed methodology aims to address this gap.

SOURCE     arXiv cs.CY (Computers and Society) (tier 1)
   ↓
DOCUMENT   277b0ec3-889a-4e9e-ba65-59d8658f5d7d
           https://arxiv.org/abs/2609.16004
   ↓
EVIDENCE   2 extracted excerpts
   ↓
MODEL RUN  5 runs, methodology 1.0.1
   ↓
SCORE      -29  (Adverse)
   ↓
CONFIDENCE 74%