Why Agents Compromise Safety Under Pressure

• URL: http://arxiv.org/abs/2603.14975v1
• Date: Mon, 16 Mar 2026 08:37:34 GMT
• Title: Why Agents Compromise Safety Under Pressure
• Abstract summary: This paper identifies a new concept called Agentic Pressure, which characterizes the endogenous tension emerging when compliant execution becomes infeasible.<n>We demonstrate that under this pressure agents exhibit normative drift where they strategically sacrifice safety to preserve utility.<n>We find that advanced reasoning capabilities accelerate this decline as models construct linguistic rationalizations to justify violation.
• Score: 6.429973383484847
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large Language Model agents deployed in complex environments frequently encounter a conflict between maximizing goal achievement and adhering to safety constraints. This paper identifies a new concept called Agentic Pressure, which characterizes the endogenous tension emerging when compliant execution becomes infeasible. We demonstrate that under this pressure agents exhibit normative drift where they strategically sacrifice safety to preserve utility. Notably we find that advanced reasoning capabilities accelerate this decline as models construct linguistic rationalizations to justify violation. Finally, we analyze the root causes and explore preliminary mitigation strategies, such as pressure isolation, which attempts to restore alignment by decoupling decision-making from pressure signals.

Related papers

• Agentic Pressure: The Endogenous Entropy of Reliable Autonomy [5.721126979967714]
  Agentic Pressure is a kinetic force that spontaneously emerges when the cost of compliance conflicts with the imperative of goal achievement.<n>We propose a theoretical framework that formalizes Agentic Pressure as the ratio between the required work to overcome environmental friction and the remaining capacity of the agent.<n> Empirical experiments validate this framework and show that aligned agents spontaneously compromise safety to preserve autonomy under high-pressure conditions.
  arXiv  Detail & Related papers  (2026-09-05T09:39:18Z)
• Polished but Unresolved: Identifying Late-Stage Pressure States in Long-Horizon Tool-Use Agents [21.24035103429806]
  We study late-stage pressure states, in which an agent is biased toward submitting a final answer that appears complete and polished while key constraints remain unresolved.<n>We propose Probe-Sensed Pressure Relief (PSPR), a plugin that applies lightweight pressure relief direction under moderate pressure and moves to structured organization under high pressure risk.
  arXiv  Detail & Related papers  (2026-09-01T07:24:53Z)
• Constitutional On-Policy Safe Distillation [64.02536207601112]
  On-policy self-distillation (OPSD) has emerged as an efficient post-training paradigm by using a teacher conditioned on privileged information to provide dense token-level supervision.<n>We propose Constitutional On-Policy Safe Distillation (COPSD), which first calibrates the teacher through a Cross-SFT cold-start and then performs constitution-conditioned on-policy distillation.
  arXiv  Detail & Related papers  (2026-06-02T03:17:56Z)
• Provably Secure Agent Guardrail [89.79561918065122]
  Existing defense architectures rely on empirical semantic guardrails and probabilistic large model adjudicators.<n>This paper proposes a new security paradigm for agents based on the fundamental limitations of logical reasoning.
  arXiv  Detail & Related papers  (2026-05-28T02:12:41Z)
• Asymmetric Goal Drift in Coding Agents Under Value Conflict [0.0]
  We show that GPT-5 mini, Haiku 4.5, and Grok Code Fast 1 exhibit asymmetric drift when their system prompt opposes strongly-held values like security and privacy.<n>Even strongly-held values like privacy show non-zero violation rates under sustained environmental pressure.<n>These findings reveal that shallow compliance checks are insufficient and that comment-based pressure can exploit model value hierarchies to override system prompt instructions.
  arXiv  Detail & Related papers  (2026-03-03T19:13:12Z)
• From Shallow to Deep: Pinning Semantic Intent via Causal GRPO [40.89749712474356]
  We propose Two-Stage Causal-GRPO, a framework designed to achieve intent pinning.<n>We show that TSC-GRPO significantly outperforms baselines in defending against jailbreak attacks while preserving general utility.
  arXiv  Detail & Related papers  (2026-03-03T07:02:20Z)
• VIGIL: Defending LLM Agents Against Tool Stream Injection via Verify-Before-Commit [44.24310459184061]
  LLM agents operating in open environments face escalating risks from indirect prompt injection.<n>We propose textbfVIGIL, a framework that shifts the paradigm from restrictive isolation to a verify-before-commit protocol.
  arXiv  Detail & Related papers  (2026-01-09T12:19:49Z)
• How Brittle is Agent Safety? Rethinking Agent Risk under Intent Concealment and Task Complexity [55.441602598245744]
  Current safety evaluations for LLM-driven agents primarily focus on atomic harms, failing to address sophisticated threats where malicious intent is concealed or diluted within complex tasks.<n>We address this gap with a two-dimensional analysis of agent safety brittleness under the pressures of intent concealment and task complexity.<n>Our findings reveal two critical phenomena: safety alignment degrades sharply and predictably as intent becomes obscured, and a "Complexity Paradox" emerges, where agents seem safer on harder tasks only due to capability limitations.
  arXiv  Detail & Related papers  (2025-11-11T17:27:27Z)
• When Models Outthink Their Safety: Mitigating Self-Jailbreak in Large Reasoning Models with Chain-of-Guardrails [74.63933201261595]
  Large Reasoning Models (LRMs) demonstrate remarkable capabilities on complex reasoning tasks.<n>LRMs remain vulnerable to severe safety risks, including harmful content generation and jailbreak attacks.<n>We propose the Chain-of-Guardrail (CoG), a training framework that recomposes or backtracks unsafe reasoning steps.
  arXiv  Detail & Related papers  (2025-10-24T09:32:25Z)
• Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning? [68.82210578851442]
  We investigate why safety alignment fails in reasoning models through a mechanistic interpretability lens.<n>Using a linear probing approach to trace refusal intentions across token positions, we discover a phenomenon termed as textbfrefusal cliff<n>We propose textbfCliff-as-a-Judge, a novel data selection method that identifies training examples exhibiting the largest refusal cliff to efficiently repair reasoning models' safety alignment.
  arXiv  Detail & Related papers  (2025-10-07T15:32:59Z)
• Towards Safe Reasoning in Large Reasoning Models via Corrective Intervention [53.25106308403173]
  We show that existing methods overlook the unique significance of safe reasoning, undermining their trustworthiness and posing potential risks in applications if unsafe reasoning is accessible for and exploited by malicious users.<n>We propose Intervened Preference Optimization (IPO), an alignment method that enforces safe reasoning by substituting compliance steps with safety triggers and constructing pairs for preference learning with strong signals.
  arXiv  Detail & Related papers  (2025-09-29T07:41:09Z)
• AdvChain: Adversarial Chain-of-Thought Tuning for Robust Safety Alignment of Large Reasoning Models [62.70575022567081]
  We propose AdvChain, an alignment paradigm that teaches models dynamic self-correction through adversarial CoT tuning.<n>Our work establishes a new direction for building more robust and reliable reasoning models.
  arXiv  Detail & Related papers  (2025-09-29T04:27:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.