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

EU Regulatory Changes

371 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: NeuroArmor: Safe-Variant-Guided Representation Consistency for Selective Re-Anchoring in Jailbreak Defense
This paper, published on arXiv, introduces NeuroArmor, a novel technical framework designed to defend large language models (LLMs) against "jailbreak" attacks—prompts that trick AI into generating ...
Read analysis →
arXiv: FORGE: Multi-Agent Graduated Exploitation and Detection Engineering
This document is a pre-print research paper, not a binding regulatory change. It introduces a proposed technical framework called FORGE, which stands for Multi-Agent Graduated Exploitation and Dete...
Read analysis →
arXiv: Signals and Spoils: Speculative Oracle Extractable Value in the Era of Cross-Chain Interoperability
This paper, published on arXiv, presents a speculative analysis of a novel form of financial risk in blockchain systems, termed "Speculative Oracle Extractable Value" (SOEV). It examines how cross-...
Read analysis →
arXiv: A Hybrid Approach For Malware Classification Using Secondary Features Fusion
1. This publication presents a novel hybrid approach for malware classification that fuses secondary features from multiple data sources to improve detection accuracy. While not a regulatory change...
Read analysis →
arXiv: FlowGuard: Flow Matching for Identity-Independent Detection of Data-Free Model Stealing Attacks on Energy Syst...
This publication introduces FlowGuard, a novel detection method for model stealing attacks targeting intrusion detection systems (IDS) used in energy infrastructure. The paper presents a flow-match...
Read analysis →
arXiv: Selective Token-Level Cryptographic Redaction for Privacy-Preserving Clinical Deployment of Large Language Models
This paper, published on arXiv, introduces a novel technical method for selectively redacting individual tokens—such as patient names or diagnoses—within large language model outputs using cryptogr...
Read analysis →
arXiv: Agent libOS: A Library-OS-Inspired Runtime for Long-Running, Capability-Controlled LLM Agents
arXiv: AI Agents Enable Adaptive Computer Worms
arXiv: PURGE: Projected Unlearning via Retain-Guided Erasure
arXiv: Collision Resistance of Single-Layer Neural Nets
arXiv: From Control Boundary to Insurance Claim: Reconstructing AI-Mediated Losses Through the CER Framework
arXiv: $π$Creds: Privately Inferred Credentials
arXiv: Same Weights, Different Robot: A Deployment Safety View of VLA Policies
arXiv: Don't Trust Us: A privacy-by-design android malware detection pipeline
arXiv: Ghost: Plausible Yet Unlearnable Trajectories via On-Manifold Substitution for Next-POI Privacy
arXiv: Designing a Hardware Reverse Engineering Course: Lessons from Eight Years in a Rapidly Evolving Tech Domain
arXiv: Black-box, Adaptive, Efficient, Transferable, Harmful, Applicable... Attacks Are All You Need to Break LLMs
arXiv: Q-FE: A Quantum-Native 6G Far-Edge Architecture Securing Industrial IoT Digital Twins via CSIDH-PQC and Asynch...
arXiv: Testing LLM Arithmetic Reasoning Generalization with Automatic Numeric-Remapping Attacks
arXiv: Channel Chart Location Privacy Based on Geo-Indistinguishability