Every entry below was pulled directly from a regulator’s feed in the last few days — CELLAR, EBA, ESMA, ENISA, BSI, BaFin, ANSSI and 14 other sources — classified against 21 EU compliance frameworks and summarised. Updates daily.
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Security and Privacy in the Musical Metaverse: Threat Analysis and Design Implications
This publication is a research paper, not a regulatory mandate, but it signals a critical emerging risk area for compliance teams. The paper provides a structured threat analysis of the "musical metaverse," identifying…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: The Native-Signature Boundary in Post-Quantum Distributed Authorization
A new technical paper, published on arXiv, proposes a formal framework for defining and enforcing the boundary between native cryptographic signatures and post-quantum (PQ) signatures within distributed authorization…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Preprocessing Failure and Adversarial Detection in Depthwise-Separable Edge Vision Systems
A new preprint, arXiv:2609.03453v1, details a critical vulnerability in depthwise-separable convolutional neural networks used in edge vision systems, such as those in smart cameras, drones, and industrial IoT devices.…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Privacy, Robustness, and Fairness Trade-offs in Federated Intrusion Detection: Geometric Indistinguishability at the Aggregation Interface
This publication, dated September 2026, is a technical research paper, not a binding regulation. It analyzes the inherent tensions between privacy, robustness, and fairness when applying federated learning to intrusion…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Spruce: Scalable Private Outsourced Retrieval Using Compact Embeddings
A new academic paper, titled Spruce: Scalable Private Outsourced Retrieval Using Compact Embeddings, has been published on arXiv. The paper proposes a technical method for performing private, outsourced data…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Long-Range Indirect Control-Flow Prediction in Stripped Binaries via Dual Virtual Hubs and Multi-Task Graph Learning
A new academic paper, published on arXiv on September 3, 2026, introduces a machine-learning technique that predicts indirect control-flow targets in stripped binary executables. Stripped binaries lack symbolic…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: After Cheap Discovery: From unknown to known-and-unfixed
The publication is a research paper, not a new regulation, but it carries significant implications for AI safety compliance. Titled After Cheap Discovery: From unknown to known-and-unfixed, it examines the gap between…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Memetic Search for Supersingular Elliptic Curves over $\mathbb{F}_p$
A new academic paper, published on arXiv on September 3, 2026, proposes a memetic search algorithm to efficiently find supersingular elliptic curves over finite fields. This is a mathematical optimization technique, not…
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Trust Me, I'm Your Developer: Self-Issued Authentication in Large Language Models
This publication, dated September 2026, is a technical research paper from arXiv, not a binding regulatory rule. It proposes a framework called self-issued authentication for large language models, where an AI system…
AI_SAFETYarxiv_cscr2 Sept 2026
arXiv: Beyond Small Patches: Black-Box Detection and Purification of Diverse Backdoor Triggers
A new academic paper proposes a method for detecting and removing backdoor triggers in AI models without requiring internal access to the model’s architecture. Backdoor attacks involve hidden patterns that cause an AI…
BREACHransomwarelive3 Sept 2026
Unverified leak-site claim: Storm names Chicago Partners Wealth Advisors (US) — Financial Services
BREACHransomwarelive3 Sept 2026
Unverified leak-site claim: Storm names Macquarrie (AU) — Financial Services
BREACHransomwarelive3 Sept 2026
Unverified leak-site claim: Storm names SITES Medical (US) — Healthcare
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: SENTINEL-RL: Offloading Topological Reasoning from LLM Agents in the Security Operations Center
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: A Non-Formulable Theorem: A Fundamental Limit of Finite Syntactic Systems and Its Consequences for Security and AI
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: PatchBench: Evaluating AI Agents for Vulnerability Patching
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: AI-Assisted Design of a Post-Quantum Cryptographic Accelerator: A Deployed-Silicon Case Study
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: A Black Box for Agentic Processes: Blockchain-Anchored Evidence for AI Agent Communication, Human Oversight, and GRC Audits
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Barnacle: Adaptive Multi-Leader Scheduling for DAG-Based Consensus
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Practice Makes (Im)Perfect: A Look Back at Benchmarking Practices for Microarchitectural Side-Channel Attacks
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: A Blind Trust, the Bloody Thrust: When Attacker-Controlled Hook Updates Steer AI Agent Harnesses towards Malicious Behaviors
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: NACRE: Rethinking Confidential Containers through Native Architectural Support
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Flip, Don't Shuffle: Watermarking LLMs at the Speed of Inference
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Supersingular Elliptic Curves Without Inseparable Small Degree Endomorphisms
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Inferring Hidden User Models from the Behavior of Personalized LLM Agents
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Beyond the Trust Boundary: A Critical Reassessment of the FIDO2 Threat Model
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: Rent-a-RAG: Embedding-Space Watermarks for Auditing Third-Party RAG
AI_SAFETYarxiv_cscr3 Sept 2026
arXiv: AlcaTRAz - Anchored Tree-Rule Defense Against Jailbreaks
BREACHransomwarelive3 Sept 2026
Unverified leak-site claim: Panzer names Dinas Komunikasi dan Informatika (ID) — Government & Defense
BREACHransomwarelive3 Sept 2026
Unverified leak-site claim: Wallstreet names America’s Food Basket (US) — Retail & E-Commerce