Currently free during beta - premium features coming soon. Subscribe now to lock in early access.
AI_SAFETY

EU Regulatory Changes

1501 changes tracked across 24 compliance frameworks including DORA, NIS2, GDPR, EU AI Act, Cyber Resilience Act, and more.

All DORA NIS2 GDPR CSRD MaRisk ISO27001 EU_AI_ACT CRA DSA DMA eIDAS2 SOC2 PCI_DSS HIPAA ISO42001 AMLD6 PSD3 DATA_ACT GPSR CER EUDR CVE BREACH AI_SAFETY
arXiv: Leaky Language Models: Stealing Architecture and Inference Optimizations via Per-Token Timing
arXiv: Security Vulnerability Patterns in AI-Generated Code: A Cross-Model Comparative Study
arXiv: Evaluating Large Language Models for Symbolic Security Protocol Analysis
arXiv: Buzz to Boom: Detecting Message Progression Vulnerabilities in Electron Applications via Segmented Directed Fu...
arXiv: Enhancing Attack Detection Capabilities in BACnet/IP Networks Using Machine-Learning Models
arXiv: Pure-DP Statistical Query Release at the Conjectured Square-Root Rate
This paper, published on arXiv, presents a new theoretical method for releasing statistical queries from a dataset while achieving pure differential privacy at the conjectured square-root rate. Thi...
Read analysis →
arXiv: Constant-time decoding of Gabidulin codes and their generalizations with application to RQC
This publication from July 2026 presents a new cryptographic algorithm for constant-time decoding of Gabidulin codes, which are a type of error-correcting code used in post-quantum cryptography. Th...
Read analysis →
arXiv: Chained Attacks on Drone-Based Federated Learning: From Network Disruption to Device Impersonation
This paper, published on arXiv on July 22, 2026, presents a new vulnerability analysis for drone-based federated learning systems. It demonstrates a chained attack methodology where an adversary ca...
Read analysis →
arXiv: The Ethics of Autonomous AI Agents for Offensive Security
This paper, published on arXiv, presents a detailed ethical analysis of deploying autonomous AI agents for offensive cybersecurity operations. It does not represent a regulatory change from a gover...
Read analysis →
arXiv: Small, Free, and Effective: Orchestrating Open-Weight Small Language Models to Outperform Single LLM for Malwa...
A new preprint from arXiv, published on July 22, 2026, demonstrates that orchestrating multiple open-weight small language models can outperform a single large language model in malware analysis ta...
Read analysis →
arXiv: ISAC-Assisted Channel Knowledge Map Generation for Physical Layer Authentication
This publication, dated July 22, 2026, introduces a novel technical framework for using Integrated Sensing and Communication (ISAC) systems to generate Channel Knowledge Maps (CKMs) for physical la...
Read analysis →
arXiv: Taming the Security-Energy Paradox: A Green AI Approach to Optimized Android Malware Detection
This paper, published on arXiv, presents a novel approach to Android malware detection that balances security effectiveness with energy efficiency, a concept termed the "security-energy paradox." I...
Read analysis →
arXiv: HijackKV: New Threat in Position-Independent KV Cache Reuse
A new research paper, "HijackKV: New Threat in Position-Independent KV Cache Reuse," published on arXiv, identifies a novel security vulnerability in large language model (LLM) inference systems. T...
Read analysis →
arXiv: JANUS: Foreseeing Latent Risk for Long-Horizon Agent Safety
This publication introduces JANUS, a novel framework designed to predict latent safety risks in AI agents operating over extended time horizons. Unlike existing safety tools that focus on immediate...
Read analysis →
arXiv: Defense Against LLM Backdoors using Critical Neuron Isolation Pruning
This paper, published on arXiv on July 22, 2026, introduces a new technical method for defending large language models (LLMs) against backdoor attacks. The technique, called Critical Neuron Isolati...
Read analysis →
arXiv: Adversarial Frontiers: Minimum-Norm Attack Ensembles for Robustness Evaluation
arXiv: Know Your Agent: Reconnaissance-Driven Pentesting of AI Agents
arXiv: DARWIN: Evolving Jailbreak Adversary and Guardrail for LLM Safety Evaluation and Protection
arXiv: An Automated Framework for Extracting Reachable Attack Chains from Cyber Threat Intelligence Reports
arXiv: GhostPrompt: Cross-Image Adversarial Prompt for Vision-Language Models