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

EU Regulatory Changes

571 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: PolicyGuard: Towards Test-time and Step-level Adversary Defense for Reinforcement Learning Agent
arXiv: LNTest: A Testbed for Evaluating Bitcoin Lightning Network-Based Botnets
arXiv: A Privacy-Preserving Framework Using Remote Data Science for Inter-Institutional Student Retention Prediction
arXiv: Semantic Identification of IoT Devices from Behavioral Primitives
arXiv: Detecting Functional Memorization in Code Language Models
arXiv: PI-Hunter: Automated Red-Teaming for Exposing and Localizing Prompt Injections
arXiv: Smarter Saboteurs, Better Fixers: Scaling & Security in Linear Multi-Agent Workflows
arXiv: SMSR: Certified Defence Against Runtime Memory Poisoning in Persistent LLM Agent Systems
arXiv: Fed-FBD: Federated Functional Block Diversification for Isolation, Privacy, and Surgical Unlearning
arXiv: CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agents
arXiv: MARCIM-WG: A cyber wargame proposal based on math modeling applied in a naval scenario
This document is not a regulatory change but a research paper proposing a new cyber wargame framework called MARCIM-WG, published on arXiv. It uses mathematical modeling to simulate cyber attacks a...
Read analysis →
arXiv: ECYSAP EYE: From Cyber Situational Awareness to Mission-Centric Decision Support for Enhanced Cyberspace Opera...
This publication, titled ECYSAP EYE, presents a research framework for integrating cyber situational awareness with mission-centric decision support, specifically aimed at enhancing cyberspace oper...
Read analysis →
arXiv: OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agents
As a senior EU regulatory compliance analyst, I summarize the following regulatory-relevant publication for compliance professionals. This paper, OCELOT, introduces a new framework for measuring a...
Read analysis →
arXiv: A Five-Plane Reference Architecture for Runtime Governance of Production AI Agents
A new technical paper published on arXiv proposes a five-plane reference architecture for runtime governance of production AI agents, titled A Five-Plane Reference Architecture for Runtime Governan...
Read analysis →
arXiv: Selection Integrity for LLM Graph Memory: An Accumulability Criterion for Information-Flow-Blind Retrieval
This paper, published on arXiv under the AI Safety framework, introduces a new technical criterion called "accumulability" for evaluating the integrity of information retrieval from large language ...
Read analysis →
arXiv: Partitioned Tags, Shared Data: Reconciling Strict Cache Isolation with Write-Shared Coherence
This publication from arXiv, dated June 10, 2026, presents a novel hardware architecture approach titled "Partitioned Tags, Shared Data." It proposes a method to reconcile strict cache isolation wi...
Read analysis →
arXiv: Reinforcement Learning Disrupts Gradient-Based Adversarial Optimization
This publication presents a research paper demonstrating that reinforcement learning (RL) can effectively circumvent standard gradient-based adversarial attacks used to test AI system robustness. T...
Read analysis →
arXiv: Bridging the Smart City Cybersecurity Data Gap Through AI-Driven Synthetic Dataset Generation
This paper, published on arXiv on June 10, 2026, proposes a novel AI-driven framework for generating synthetic datasets to address critical data-sharing gaps in smart city cybersecurity. The author...
Read analysis →
arXiv: Mind your key: An Empirical Study of LLM API Credential Leakage in iOS Apps
A new empirical study published on arXiv, titled "Mind your key: An Empirical Study of LLM API Credential Leakage in iOS Apps," reveals a systemic vulnerability in mobile applications that integrat...
Read analysis →
arXiv: Categorical Robustness Assessment for Machine Learning based Network Intrusion Detection Systems
This publication introduces a new methodology for assessing the categorical robustness of machine learning models used in network intrusion detection systems. It proposes a framework that evaluates...
Read analysis →