Tag: agentic AI security
Browse our exclusive articles!
FedProxy: Advanced Federated Fine-Tuning for Large Language Models
Discover FedProxy, a cutting-edge framework enabling secure, efficient federated fine-tuning of LLMs with privacy and performance across diverse data.
Hierarchically Robust Zero-Shot Vision-Language Models
Discover how hierarchical embeddings improve adversarial robustness in zero-shot vision-language models for enhanced performance and security.
Soft-Failure Attacks on Retrieval-Augmented Generation Explained
Discover how soft-failure attacks undermine Retrieval-Augmented Generation (RAG) systems and the innovative DEJA method challenging AI security.
Owner-Harm: Key AI Safety Threat to Deployers
Explore the Owner-Harm threat model highlighting AI risks to deployers and new defense strategies to improve AI agent safety and security.
How Adversarial Environments Exploit Agentic AI Flaws
Discover how adversarial environments deceive agentic AI, exposing trust gaps and vulnerabilities in tool-integrated systems.