A Closer Look at the Existing Risks of Generative AI: Mapping the Who, What, and How of Real-World Incidents

2025-05-28

AI Summary

This study investigates the risk evolution mechanisms and cross-stakeholder harm propagation pathways of generative AI in real-world settings. Method: Drawing on empirical analysis of 499 publicly documented incidents, we construct the first domain-specific failure taxonomy for generative AI and propose a three-dimensional “risk–failure–stakeholder” mapping framework. Using systematic literature review, incident coding, and co-occurrence statistics, we quantify prevalent failure patterns, associated harm types, and affected stakeholder distributions. Results: We reveal a defining characteristic distinguishing generative AI from traditional AI: over 70% of incidents originate in usage contexts, yet harms substantially externalize to non-primary users—including the general public, environment, and vulnerable populations. The study establishes the primacy of non-technical governance interventions and provides empirically grounded foundations for policy design, developer accountability frameworks, and public risk literacy initiatives.

Technology Category

• Philosophy and Ethics of AI: Artificial General Intelligence
• Humans and AI: Learning Human Values and Preferences
• Reasoning under Uncertainty: Causality

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization

Abstract

Due to its general-purpose nature, Generative AI is applied in an ever-growing set of domains and tasks, leading to an expanding set of risks of harm impacting people, communities, society, and the environment. These risks may arise due to failures during the design and development of the technology, as well as during its release, deployment, or downstream usages and appropriations of its outputs. In this paper, building on prior taxonomies of AI risks, harms, and failures, we construct a taxonomy specifically for Generative AI failures and map them to the harms they precipitate. Through a systematic analysis of 499 publicly reported incidents, we describe what harms are reported, how they arose, and who they impact. We report the prevalence of each type of harm, underlying failure mode, and harmed stakeholder, as well as their common co-occurrences. We find that most reported incidents are caused by use-related issues but bring harm to parties beyond the end user(s) of the Generative AI system at fault, and that the landscape of Generative AI harms is distinct from that of traditional AI. Our work offers actionable insights to policymakers, developers, and Generative AI users. In particular, we call for the prioritization of non-technical risk and harm mitigation strategies, including public disclosures and education and careful regulatory stances.

Problem

• Identifies real-world risks and harms from Generative AI applications
• Analyzes 499 incidents to map failure modes and impacted stakeholders
• Highlights need for non-technical mitigation strategies and policy action

Innovation

• Constructed taxonomy for Generative AI failures
• Systematically analyzed 499 reported incidents
• Prioritized non-technical risk mitigation strategies