Invisible failures in human-AI interactions

2026-03-16

AI Summary

This study addresses the challenge of “invisible failures” in AI systems—errors that occur during human-AI interaction without user awareness or explicit feedback, rendering them difficult to detect and mitigate. Leveraging the large-scale WildChat dialogue dataset, the work introduces the first taxonomy of invisible failures, comprising eight prototypical categories that distinguish between interaction-driven and capability-driven failure modes. Through clustering, behavioral coding, and cross-domain quantitative analysis, the authors uncover underlying causes and co-occurrence patterns. Their findings reveal that 78% of AI failures are invisible, with 91% stemming from flaws in interaction design, and 94% projected to persist even with more capable models. This framework establishes a systematic foundation for monitoring, analyzing, and improving AI reliability for developers, researchers, and policymakers alike.

Technology Category

• Humans and AI: Interaction Techniques and Devices
• Philosophy and Ethics of AI: Accountability, Interpretability & Explainability
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

AI systems fail silently far more often than they fail visibly. In a large-scale quantitative analysis of human-AI interactions from the WildChat dataset, we find that 78% of AI failures are invisible: something went wrong but the user gave no overt indication that there was a problem. These invisible failures cluster into eight archetypes that help us characterize where and how AI systems are failing to meet users' needs. In addition, the archetypes show systematic co-occurrence patterns indicating higher-level failure types. To address the question of whether these archetypes will remain relevant as AI systems become more capable, we also assess failures for whether they are primarily interactional or capability-driven, finding that 91% involve interactional dynamics, and we estimate that 94% of such failures would persist even with a more capable model. Finally, we illustrate how the archetypes help us to identify systematic and variable AI limitations across different usage domains. Overall, we argue that our invisible failure taxonomy can be a key component in reliable failure monitoring for product developers, scientists, and policy makers. Our code and data are available at https://github.com/bigspinai/bigspin-invisible-failure-archetypes

Problem

• invisible failures
• human-AI interactions
• failure taxonomy
• interactional dynamics
• AI reliability

Innovation

• invisible failures
• failure archetypes
• human-AI interaction
• interactional dynamics
• AI reliability