2,077+ academic papers on AI security, attacks, and defenses
Showing 1–20 of 51 papers
Clear filtersCharoes Huang, Xin Huang, Amin Milani Fard
Prompt injection is listed as the number-one vulnerability class in the OWASP Top 10 for LLM Applications that can subvert LLM guardrails, disclose...
Octavian Untila
An autonomous AI ecosystem (SUBSTRATE S3), generating product specifications without explicit instructions about formal methods, independently...
Uchi Uchibeke
AI agents today have passwords but no permission slips. They execute tool calls (fund transfers, database queries, shell commands, sub-agent...
Vincent Siu, Jingxuan He, Kyle Montgomery +4 more
Security in LLM agents is inherently contextual. For example, the same action taken by an agent may represent legitimate behavior or a security...
Md Takrim Ul Alam, Akif Islam, Mohd Ruhul Ameen +2 more
Large language models (LLMs) deployed behind APIs and retrieval-augmented generation (RAG) stacks are vulnerable to prompt injection attacks that may...
Taiwo Onitiju, Iman Vakilinia
Large Language Models increasingly power critical infrastructure from healthcare to finance, yet their vulnerability to adversarial manipulation...
Zhouwei Zhai, Mengxiang Chen, Anmeng Zhang
Large language models offer transformative potential for e-commerce search by enabling intent-aware recommendations. However, their industrial...
Cosimo Spera
Customer service automation is undergoing a structural transformation. The dominant paradigm is shifting from scripted chatbots and single-agent...
Yihao Zhang, Zeming Wei, Xiaokun Luan +7 more
Autonomous LLM-based agents increasingly operate as long-running processes forming densely interconnected multi-agent ecosystems, whose security...
Yihao Zhang, Zeming Wei, Xiaokun Luan +7 more
Autonomous LLM-based agents increasingly operate as long-running processes forming densely interconnected multi-agent ecosystems, whose security...
Zhuoshang Wang, Yubing Ren, Yanan Cao +3 more
While watermarking serves as a critical mechanism for LLM provenance, existing secret-key schemes tightly couple detection with injection, requiring...
Ziling Zhou
AI agents dynamically acquire tools, orchestrate sub-agents, and transact across organizational boundaries, yet no existing security layer verifies...
Ziling Zhou
AI agents dynamically acquire capabilities at runtime via MCP and A2A, yet no framework detects when capabilities change post-authorization. We term...
Jiangrong Wu, Zitong Yao, Yuhong Nan +1 more
Tool-augmented LLM agents increasingly rely on multi-step, multi-tool workflows to complete real tasks. This design expands the attack surface,...
Sarbartha Banerjee, Prateek Sahu, Anjo Vahldiek-Oberwagner +2 more
Rapid progress in generative AI has given rise to Compound AI systems - pipelines comprised of multiple large language models (LLM), software tools...
Frank Li
Tool-augmented LLM agents introduce security risks that extend beyond user-input filtering, including indirect prompt injection through fetched...
Chingkwun Lam, Jiaxin Li, Lingfei Zhang +1 more
Long-term memory has emerged as a foundational component of autonomous Large Language Model (LLM) agents, enabling continuous adaptation, lifelong...
Raj Sanjay Shah, Jing Huang, Keerthiram Murugesan +2 more
Unlearning in Large Language Models (LLMs) aims to enhance safety, mitigate biases, and comply with legal mandates, such as the right to be...
Xiangwen Wang, Ananth Balashankar, Varun Chandrasekaran
Large language models remain vulnerable to jailbreak attacks, yet we still lack a systematic understanding of how jailbreak success scales with...
Zixun Xiong, Gaoyi Wu, Lingfeng Yao +3 more
Communication topology is a critical factor in the utility and safety of LLM-based multi-agent systems (LLM-MAS), making it a high-value intellectual...
Get breaking CVE alerts, compliance reports (ISO 42001, EU AI Act), and CISO risk assessments for your AI/ML stack.
Start 14-Day Free Trial
AI Threat Alert