CS Paper Radar

When AI Agents Act on Our Behalf: A Taxonomy of Human-Agent Interaction Failure

Canwen Wang, Qing Xiao, Haiyi Zhu, Hong Shen

cs.HC cs.AIcs.CY ·submitted 2026-10-30 19:17 UTC · v1

Abstract

AI agents write and execute code, manipulate user interfaces, and transact on their users' behalf. We ask what forms failure takes in human-agent interaction, and how those forms change as agents act with greater autonomy. From 695 documented incidents in two public repositories, we developed 12 failure modes bottom-up within three interaction stages adapted from Amershi et al.'s guidelines, coding the autonomy at which each arose. What a person can do when an agent fails depends on how much the agent completes between points of user involvement--a property fixed at deployment. Therefore, the same model on the same task can fail differently under different configurations. All 12 modes appear at both low and high autonomy, and rising autonomy changes the form each failure takes. Users need controls between full delegation and full manual work, and neither incident records nor agent benchmarks reach the interaction where these failures occur.

Closest papers

Nearest neighbours by meaning across the whole index.

Babysitting AI Agents: A User-Centered Taxonomy and Study of Agentic Delegation Failures

cs.HC cs.AI ·2026-10-26 ·Shravika Mittal, Sunnie S. Y. Kim, Scott Counts, Jennifer Neville, Debarati Das +1 more
similarity 0.63
Abstract

Users are increasingly delegating tasks to AI agents, allowing agents to perform long-running, multi-step workflows on their behalf. However, agents may perform these tasks in ways that depart from users' expectations for how they should be carried out, leading to what we call agentic delegation failures. We leverage a large corpus of 163,487 Reddit posts, describing AI agent use, to develop MANDATE: a taxonomy of 23 failure categories organized into seven groups. A follow-up study with 73 AI agent users examined the relationship between prevalence and perceptions of severity and noticeability, finding that failures rated as more severe are considered harder to notice, and that prior experience with a failure is associated with greater noticeability. Finally, we assess the maturity of research on each failure and identify commonly encountered failures that have received little attention (e.g., agents running without making progress). Our findings inform opportunities to make delegation easier for users to oversee and better studied by the research community.

0.13

Why Do Agent Teams Fail Across Boundaries? A Failure Taxonomy for Agent-to-Agent Systems with Private States

cs.MA cs.AI ·2026-10-14 ·Sarath Shekkizhar, Prathyusha Jwalapuram, Zixuan Ke, Vaidehi Patil, Jiayu Wang +10 more
code/data similarity 0.63
Abstract · 20 pages, 7 figures, 4 tables. Run logs: https://huggingface.co/datasets/a2a-salesforce/a2-failures

Agents are increasingly interacting with other agents that act for different principals (people or organizations on whose behalf they work), such as a buyer's personal agent negotiating with an enterprise's seller agent. In such agent-to-agent (A2A) scenarios, each agent has its own environment and holds information and authority that the other cannot see. Moreover, the interests of each agent may be aligned (cooperative), partly aligned (mixed-motive), or opposed (competitive). Existing multi-agent system failure studies assume one orchestrator, shared workspace, or single principal. We ask how A2A systems fail when these assumptions are removed. We study more than 2,500 recorded runs of A2A systems built from different agent harnesses and models, spanning 488 scenarios across eight families (legal, CRM, security, commerce, etc.). From these traces, we derive $A^2$-Failures, a taxonomy of 20 A2A failure modes in four categories: (i) discovery and delegation, (ii) grounding and record consistency, (iii) verification, and (iv) conduct and fidelity. Three findings hold across every harness, model, and scenario studied: (1) discovery and interaction are driven by the phrasing of the task rather than the availability of peer agents; (2) no agent verifies a peer's factual claim on its own initiative, so an A2A system inherits the reliability of its least reliable participant unless the protocol or mechanism performs the check; and (3) the aggregate outcome score credits the agent completing the task while staying blind to the coordination, verification, and conduct failures beneath it, so A2A systems must be evaluated per agent, against each agent's own principal. We release the run logs behind the taxonomy at https://huggingface.co/datasets/a2a-salesforce/a2-failures.

0.34

Risks and Controls for Multi-Agent Systems: an analytical framework for deployment of AI agents across organisational boundaries

cs.MA cs.AI ·2026-08-27 ·Alistair Reid, Simon O'Callaghan, Dustin Venini, Liam Carroll, Tiberio Caetano
similarity 0.63
Abstract · This paper has been published by the Australian AI Safety Institute within the Department of Industry, Science and Resources under a CC BY 4…

This report presents a framework to help organisations, policymakers and researchers reason about the risks that emerge when AI agents interact with each other, how those risks change as interactions cross organisational boundaries, and the controls that may help address them. As organisations deploy AI agents, those agents will increasingly interact with each other: inside the organisation, with the agents of partners, customers and suppliers, and with unknown counterparties on the open internet. Failures can emerge from the interactions themselves, and once those interactions cross an organisation's perimeter, no single organisation can fully see, control or govern them. The report introduces three deployment tiers, defined by the minimum common governance binding any two interacting agents: singular governance, where one organisation governs every agent; federated governance, where multiple organisations deploy into a shared environment under agreed rules; and open environments, where agents operate with no central authority and shared standards are adopted voluntarily if at all. Within each tier, the report examines risk factors, failure modes and available controls. It identifies who is positioned to apply the controls, and where no actor is positioned to act, it characterises the gap and the collective action required to close it.

0.34

A Competing-Hazards Systematization of Loss of Control in Autonomous Agents

cs.AI cs.CR ·2026-09-29 ·Mohamed Aly Bouke
similarity 0.62
Abstract · 12 pages, 1 figure, 3 tables. Dataset: https://doi.org/10.5281/zenodo.22995268

Leading AI developers have reported agents acting beyond their approved limits, which a United Nations panel described as an early warning of loss of human control. Yet incident reports and agent-safety evaluations describe these events differently, making it difficult to compare failures, trace risk across attempts, or separate agent behavior from the environment's role in allowing an out-of-scope action to succeed. To address this gap, we introduce a common framework in which each attempt ends in approved completion, safe stopping, scope escape, or continuation. We formalize the framework as a discrete-time competing-hazards model and derive escape probability within a retry budget, a model-conditional safe-budget limit, and conditions for estimation from execution logs. We audit 22 incident reports and 102 agent-safety evaluations published from January 2025 to September 2026 using primary sources. Six incidents involved tasks that could not be completed within scope, thirteen involved agents that continued rather than stopped, and five did not report stopping behavior. Developers' figures imply a task-level incidence ratio near 47 for out-of-scope coordination in never-solved versus solved tasks. Among evaluations, 87 recorded an out-of-scope effect or specification violation, 26 treated safe stopping as a first-class outcome, only 20 recorded both, and 79 merged budget exhaustion with failure. In 20 of 22 incidents, the environment allowed an out-of-scope effect, indicating that realized loss of control often reflected persistent agent behavior interacting with permissive boundary conditions; meanwhile, no evaluation reported all fields needed to estimate the full competing-hazards process from published evidence.

0.25

From Language Models to World-Acting Systems: Progress and Limits of Agentic AI across Digital, Social, Virtual, and Physical Environments

cs.AI cs.LGcs.MA ·2026-09-04 ·Linsen Zhu, Mengqing Cai
similarity 0.62
Abstract · Review article. 29 pages, 1 figure, 3 tables. Literature cutoff: 31 August 2026

Large language models become consequential agents when surrounding systems let outputs change external state. Models now call tools, operate interfaces, delegate work, retain state, inhabit generated worlds, and control robots or laboratory equipment. Such advances are often narrated as one march toward autonomy, conflating model competence, system integration, persistence, and safe authority. This critical review synthesizes primary research and official technical specifications available by 31 August 2026. We organize the evidence along delegated authority, temporal persistence, and environmental coupling, while separating model, harness, and environment. Within the evidence examined, action-interface expansion is documented more convincingly than robust completion, recovery, authorization, or independent verification. Model Context Protocol and Agent2Agent improve interoperability but do not establish trustworthy delegation; multi-agent organization adds specialization alongside cost and correlated failure. Persistent simulations and world models support training and planning but do not themselves demonstrate agency; robotics and self-driving laboratories establish bounded feasibility rather than unattended open-world reliability. We propose justified delegation as an analytical and normative heuristic, not an observed law or certified score: expand action scope only where evidence supports provenance, bounded authority, failure detection, safe recovery, and calibrated human control. This framing yields a research agenda for coupled model-harness evaluation, capability-based permissions, durable state, cross-agent accountability, and staged physical validation.

0.38

An Autonomy Aware Metamodel for Human AI Collaboration in Software Engineering

cs.SE ·2026-09-06 ·Md Mahade Hasan, Md Toufique Hasan, Zheying Zhang, Kalle Kulonen, Konsta Kalliokoski +3 more
similarity 0.61
Abstract

Artificial Intelligence (AI) is shifting software engineering from tool-supported processes towards AI-first collaboration, where authority is dynamically distributed across human and artificial actors. However, existing method engineering approaches assume static, human-centric control and provide limited support explicitly capturing evolving autonomy. This paper presents a vision for autonomy-aware method engineering by proposing a metamodel that treats autonomy not as a fixed property of an actor, but as a derived, situation-dependent authority assignment determined by task, context, and collaboration pattern. The metamodel formalizes autonomy through four authority dimensions: task execution, task decomposition, task initiation, and collaboration reconfiguration. Through an analytical instantiation with a multi-agent requirements analysis tool, we illustrate how the metamodel supports dynamic authority assignment. This work provides a conceptual foundation for governance-aware, adaptable, and AI-first software engineering methods.

0.28

A Mechanistic Study of Automated Failure Attribution in LLM-based Agentic Systems

cs.LG ·2026-10-20 ·Wendy Zheng, Liang Wu, Yinhan He, Mayank Darbari, Jundong Li +1 more
benchmarkcode/data similarity 0.61
Abstract

Agentic systems have demonstrated strong capabilities on complex tasks by enabling multiple interacting large language model (LLM)-based agents to collaborate and execute subtasks in parallel. However, this collaborative structure introduces new challenges for reliably evaluating and diagnosing system failures, particularly failure attribution: tracing a system's execution to identify the agent step responsible for an incorrect final outcome. Because manually performing this task is time- and labor-intensive, recent work has turned to LLMs as judges to automate the process, referred to as automated failure attribution (AFA). These approaches assume that a sufficiently capable model can infer the failing step directly from an interaction transcript. In practice, however, direct prompting often performs poorly, which motivates follow-up work that explores various ways to improve AFA performance, including alternative prompting strategies and inference-time interventions. In this paper, we examine a more fundamental question: whether this weak behavioral performance reflects a genuine capability limitation. Using linear probes applied to internal activations, we uncover a significant \emph{knows-but-won't-say gap:} models encode sufficient information to attribute a failure correctly but fail to explicitly output the responsible step. Building on this finding, we propose \emph{grafting}, a simple manual intervention that exploits knowledge already encoded in the model to improve AFA performance without requiring ground-truth labels at inference time. Extensive evaluations across multiple LLMs and benchmark settings demonstrate that grafting consistently outperforms direct-prompting approaches, underscoring the potential of leveraging latent model knowledge for more reliable failure attribution. Our code is available at https://github.com/wendyzheng22/mi4afa.

0.29

Let the Agent Do It? How Software Practitioners Understand and Make Permission Decisions in Agentic AI Assistants

cs.SE ·2026-10-05 ·Larissa Salerno, Haoyu Gao, Gregory Gay, Alexander Serebrenik, Philipp Leitner
similarity 0.61
Abstract · 21 pages, 6 figures

Agentic AI assistants increasingly act on developers' behalf by modifying files, executing commands, and accessing external resources. These actions often require permission, yet little is known about how practitioners make permission decisions while still benefiting from agent autonomy. To address this gap, we conducted a sequential mixed methods study, interviewing 18 practitioners who use AI agents and then surveying 115 practitioners based on the interview findings. We find that practitioners often understand agent behaviour through what they can directly observe and review, while decisions, data use, and other activity behind the scenes remain less clear. This uncertainty also shapes permission decisions, which depend on the scope and risk of an action, whether it fits the task, familiarity with the agent, and the environment in which it operates. Practitioners respond by adjusting how closely they oversee agents, from setting limits in advance to monitoring execution and reviewing work afterwards. How much scrutiny they apply depends on factors such as trust, task importance, time pressure, and the consequences of an action. Our findings suggest that permission systems should make consequential actions easier to review, distinguish what an agent is allowed to do from what the user intended, make reversibility clearer, avoid treating repeated approvals as stable preferences, and distinguish rejecting a single action from rejecting an entire approach.

0.15