From Thinker to Society: Security in Hierarchical Autonomy Evolution of AI Agents

2026-03-08 · Xiaolei Zhang, Lu Zhou, Xiaogang Xu, Jiafei Wu, Tianyu Du, Heqing Huang, Hao Peng, Zhe Liu arxiv

Artificial Intelligence (AI) agents have evolved from passive predictive tools into active entities capable of autonomous decision-making and environmental interaction, driven by the reasoning capabilities of Large Language Models (LLMs). However, this evolution has introduced critical security vulnerabilities that existing frameworks fail to address. The Hierarchical Autonomy Evolution (HAE) framework organizes agent security into three tiers: Cognitive Autonomy (L1) targets internal reasoning integrity; Execution Autonomy (L2) covers tool-mediated environmental interaction; Collective Autonomy (L3) addresses systemic risks in multi-agent ecosystems. We present a taxonomy of threats spanning cognitive manipulation, physical environment disruption, and multi-agent systemic failures, and evaluate existing defenses while identifying key research gaps. The findings aim to guide the development of multilayered, autonomy-aware defense architectures for trustworthy AI agent systems.

📄 PDF Abstract BibTeX arXiv:2603.07496

Code (0)

등록된 구현이 없습니다.

Cyber Autonomy: Automating the Hacker- Self-healing, self-adaptive, automatic cyber defense systems and their impact to the industry, society and national security

2020-12-08 · Ryan K L Ko

This paper sets the context for the urgency for cyber autonomy, and the current gaps of the cyber security industry. A novel framework proposing four phases of maturity for full cyber autonomy will be discussed. The pape…

Ethics

Godseed: Benevolent or Malevolent?

2014-02-01 · Eray Özkural

It is hypothesized by some thinkers that benign looking AI objectives may result in powerful AI drives that may pose an existential risk to human society. We analyze this scenario and find the underlying assumptions to b…

AI Agent Cultural Vocal Bursts Intensity Prediction

Human Society-Inspired Approaches to Agentic AI Security: The 4C Framework

2026-02-02 · Alsharif Abuadbba, Nazatul Sultan, Surya Nepal, Sanjay Jha arxiv

AI is moving from domain-specific autonomy in closed, predictable settings to large-language-model-driven agents that plan and act in open, cross-organizational environments. As a result, the cybersecurity risk landscape…

Cog-Rethinker: Hierarchical Metacognitive Reinforcement Learning for LLM Reasoning

2025-10-13 · Zexu Sun, Yongcheng Zeng, Erxue Min, Heyang Gao 외 arxiv

Contemporary progress in large language models (LLMs) has revealed notable inferential capacities via reinforcement learning (RL) employing verifiable reward, facilitating the development of O1 and R1-like reasoning mode…

Reinforcement Learning Mathematical Reasoning

TKG-Thinker: Towards Dynamic Reasoning over Temporal Knowledge Graphs via Agentic Reinforcement Learning

2026-02-05 · Zihao Jiang, Miao Peng, Zhenyan Shan, Wenjie Xu 외 arxiv

Temporal knowledge graph question answering (TKGQA) aims to answer time-sensitive questions by leveraging temporal knowledge bases. While Large Language Models (LLMs) demonstrate significant potential in TKGQA, current p…

Graph Question Answering Reinforcement Learning Knowledge Graphs