agent/papers/items/2026-2603-27148-safetydrift-predicting-when-ai-agents-cross-the-line-before-they-actually-do.md

Paper: SafetyDrift: Predicting When AI Agents Cross the Line Before They Actually Do

---

type: paper title: "SafetyDrift: Predicting When AI Agents Cross the Line Before They Actually Do" authors: Aditya Dhodapkar, Farhaan Pishori year: 2026 venue: arXiv url: https://arxiv.org/abs/2603.27148 code_url: source: arxiv collected_at: 2026-07-08 published_at: 2026-03-28 updated_at: 2026-03-28 status: queued relevance: high topics:

• agent-safety
• tool-use methods:

benchmarks:

models:

datasets:

• cs.CR
• cs.AI related_concepts:

related_jobs:

related_experiments:

related_projects:

collection_score: 14 collection_queries: agent-safety

One-line Takeaway

Auto-collected from arXiv because it matched the Agent collection queries. Needs human skim.

Why Collected

• matched queries: agent-safety
• inferred topics: agent-safety, tool-use
• arXiv categories: cs.CR, cs.AI
• collection score: 14

Review Checklist

• Does this paper directly inform Agent architecture, evaluation, memory, tools, safety, coding agents, GUI/browser agents, or multi-agent workflows?
• Does it include a benchmark, dataset, code, or reproducible experimental setup?
• Should it be promoted from queued to skimmed or summarized?

Links

• arXiv: https://arxiv.org/abs/2603.27148