Security in Multi-Agent LLM Systems
Research on security challenges and safety issues in systems where multiple LLM agents interact, including evolutionary frameworks and social simulations
Security in Multi-Agent LLM Systems
Research on Large Language Models in Security in Multi-Agent LLM Systems
AgentBreeder: Safer Multi-Agent LLM Systems
Evolutionary self-improvement framework balancing capability and safety
Combating Digital Wildfire
How LLM Agents Can Model Rumor Propagation in Social Networks
Building Responsible AI Agent Systems
Addressing Security Challenges in LLM-powered Multi-Agent Systems
Exploiting the Weak Links in LLM Agents
Security vulnerabilities in commercial LLM agent systems beyond the models themselves
AgentGuard: Security for AI Tool Systems
Automated detection and prevention of unsafe AI agent workflows
Multi-Agent LLMs for Advanced Cybersecurity
Collaborative AI agents outperforming single agents in offensive security tasks
Securing Multi-Agent LLM Systems
A topology-based framework for detecting and mitigating security threats
Exploiting LLM Agent Memory
New privacy vulnerabilities in AI assistants' memory systems
Security Vulnerabilities in LLM-Powered Agent Systems
New attack vector compromises multi-agent collaboration
Security Vulnerabilities in LLM Multi-Agent Systems
Exposing Communication Channels as Attack Vectors
Securing AI Agents Against Jailbreak Attacks
Novel system for protecting autonomous agents from multi-turn exploitation
Simulating Echo Chambers with AI Agents
Using LLMs to model social media polarization dynamics
Web AI Agents: A New Security Frontier
Understanding why web-enabled AI systems face unique vulnerabilities
Gaming the System: LLMs as Deceptive Players
A novel game-based framework to assess AI persuasion capabilities
Securing Multi-Agent AI Systems
A Hierarchical Framework for LLM-based Agent Safety
The Deception Risk in LLM Mixtures
Exposing Vulnerabilities in Collaborative AI Systems
Bot Wars: Fighting Fire with AI
Using LLMs to Counter Phone Scams Through Strategic Deception
Securing LLM Agents: The TrustAgent Framework
A comprehensive approach to identify and mitigate security threats in LLM agent systems
Multi-Agent Cooperation: Beyond Human Decision-Making
Advancing AI collaboration in complex scenarios
Securing LLM Agents Against Privilege Escalation
A novel protection mechanism for AI agent systems
Smart LLM Traffic Control
Using 'Number of Thoughts' to route prompts and detect attacks
Real-Time Video Surveillance with AI Intelligence
A breakthrough online video anomaly detection system powered by LLMs
Multi-Agent LLMs vs. Phishing Attacks
Using AI Debate to Better Detect Evolving Phishing Threats
EncGPT: Revolutionizing Communication Security
Dynamic Encryption Through Multi-Agent LLM Collaboration
Vulnerabilities in Multi-Agent LLM Systems
Breaking pragmatic systems with optimized prompt attacks
Personality-Driven AI Agents for Security
How personality traits influence autonomous LLM-based agents' decision-making
Building Safer AI Agents
Security Challenges in Foundation AI Systems
Security Vulnerabilities in Multi-Tool LLM Agents
Discovering Cross-Tool Harvesting and Polluting (XTHP) Attacks
Security Vulnerabilities in Model Context Protocol
Critical exploits found in widely-adopted AI integration standard
The Hidden Fragility of LLM Routers
Exposing security vulnerabilities in AI model routing systems
