Explorar

Noticias de IA

1359 elementos — filtrados, clasificados y sin duplicados

arXiv cs.AI Security & Safety
Black-box Membership Inference Attacks on the Pre-training Data of Image-generation Models
arXiv cs.AI Security & Safety
Cordyceps: Covert Control Attacks on LLMs via Data Poisoning
arXiv cs.AI Security & Safety
Lessons from Penetration Tests on Large-Scale Agent Systems
arXiv cs.AI Security & Safety
Jailbreak susceptibility prediction and mitigation via the behavioral geometry of models
arXiv cs.AI Security & Safety
Tracing the Dynamics of Refusal: Exploiting Latent Refusal Trajectories for Robust Jailbr…
arXiv cs.AI Security & Safety
SkillSieve: A Hierarchical Triage Framework for Detecting Malicious AI Agent Skills
arXiv cs.AI Security & Safety
Turning Bias into Bugs: Bandit-Guided Style Manipulation Attacks on LLM Judges
arXiv cs.AI Security & Safety
Furina: Fragmented Uncertainty-Driven Refusal Instability Attack
Hugging Face Daily Papers Security & Safety
MRMMIA: Membership Inference Attacks on Memory in Chat Agents
Hugging Face Daily Papers Security & Safety
Disentangling Adversarial Prompts: A Semantic-Graph Defense for Robust LLM Security
Ars Technica AI Security & Safety
Millions of AI agents imperiled by critical vulnerability in open source package
Ars Technica AI Security & Safety
FBI agent explains how easy it is to ID people posting AI porn without consent
Hugging Face Daily Papers Security & Safety
Do Modern Post-Hoc Watermarking Methods Beat Broken-Arrows?
Hugging Face Daily Papers Security & Safety
Black-box Membership Inference Attacks on the Pre-training Data of Image-generation Models
Bloomberg Technology Security & Safety
BNP Paribas Works With Mistral to Prep for Mythos-Like AI Models
arXiv cs.AI Security & Safety
Committed SAE-Feature Traces for Audited-Session Substitution Detection in Hosted LLMs
arXiv cs.AI Security & Safety
Safety in Embodied AI: A Survey of Risks, Attacks, and Defenses
arXiv cs.AI Security & Safety
Security of OpenClaw Agents: Fundamentals, Attacks, and Countermeasures
arXiv cs.AI Security & Safety
When the Manual Lies: A Realistic Benchmark to Evaluate MCP Poisoning Attacks for LLM Age…
arXiv cs.AI Security & Safety
Attested Tool-Server Admission: A Security Extension to the Model Context Protocol
arXiv cs.AI Security & Safety
Enhancing Reliability in LLM-Based Secure Code Generation
arXiv cs.AI Security & Safety
SoK: A Comprehensive Security Analysis of Jailbreak Resilience in GPT and DeepSeek Models
arXiv cs.AI Security & Safety
Explainable Attention-Guided Stacked Graph Neural Networks for Malware Detection
arXiv cs.AI Security & Safety
Membership Inference Attacks on Tokenizers of Large Language Models
arXiv cs.AI Security & Safety
LivePI: More Realistic Benchmarking of Agents Against Indirect Prompt Injection
arXiv cs.AI Security & Safety
Hidden-State Privacy Has an Empty Middle
arXiv cs.AI Security & Safety
Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-…
arXiv cs.AI Security & Safety
Uncovering Vulnerabilities of LLM-Assisted Cyber Threat Intelligence
arXiv cs.AI Security & Safety
Concept Drift Adaptation Using Self-Supervised and Reinforcement Learning In Android Malw…
arXiv cs.AI Security & Safety
Batch Normalization Amplifies Memorization and Privacy Risks