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

EU Regulatory Changes

1525 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: Quantum One-Way Functions and Related Cryptographic Primitives
arXiv: ABC: Numerical Data Collection under Local Differential Privacy without Prior Knowledge
arXiv: Algebraic Cryptanalytic Extraction on Hard-Label Neural Networks
arXiv: DreamGuard: Efficient Runtime Guardrail for LLM Agents via Risk-Aware World Model
arXiv: Breaking Customized LLMs for Coding: Automated Red Teaming for Instruction Backdoor Attacks
arXiv: Enhancing Anomaly Resilience in Research Networks: A Large-Scale Forecasting Benchmark for Dynamic Security Ba...
arXiv: Detecting Safety Training Modification in Language Models via Activation Analysis
arXiv: When Experience Becomes Instruction: Trajectory Poisoning in Self-Evolving Agent Skill Systems
arXiv: A Self-Explainable Deep Architecture for Security Applications
arXiv: Behavioral Residualization for Unsupervised Intrusion Detection in Automotive CAN Networks
arXiv: PromptShield Home: Ambient Multimodal Prompt Injection Defense for Smart-Home Agents
arXiv: Exploring Privacy Leakage and Data Disclosure Violations in the MacOS Application Ecosystem
arXiv: Agent Against Agent: An Agentic System for Automatic Prompt Injection Red Teaming
A new research paper, published on arXiv, introduces an automated system designed to test AI models for vulnerabilities to prompt injection attacks. The system, called an agentic red teaming framew...
Read analysis →
arXiv: Hardware Design and Security in the Era of Chiplets and LLMs
This publication, dated August 5, 2026, is a technical research paper from arXiv, not a binding regulation. It examines the convergence of two hardware trends: the shift to modular chiplet-based se...
Read analysis →
arXiv: Kerckhoffs-Compliant Watermarking for Physical Design IP Protection: From Placement to Routing
A new academic paper proposes a watermarking technique for protecting the intellectual property (IP) of physical chip designs, specifically targeting the entire design flow from placement to routin...
Read analysis →
arXiv: Gradient Immunity: Null-Space Resistance to Malicious Fine-Tuning
A new technical paper, Gradient Immunity: Null-Space Resistance to Malicious Fine-Tuning, has been published on arXiv. The paper proposes a method to make large language models resistant to harmful...
Read analysis →
arXiv: Private Direct Preference Optimization for LLM Alignment
On August 5, 2026, a new research paper was published on arXiv proposing a method called Private Direct Preference Optimization (DPO) for aligning large language models (LLMs) with human preference...
Read analysis →
arXiv: When Do PEFT Adaptations Leak Structure? Measuring Black-Box Structural Bounds in Public-Base Model Services
A new research paper, published on arXiv, examines how parameter-efficient fine-tuning (PEFT) methods used to adapt large language models can inadvertently leak structural information about the und...
Read analysis →
arXiv: Towards Decentralized Searcher Competition in MEV Markets
This publication, dated August 2026, is a research paper from arXiv proposing a framework to decentralize competition among searchers in Maximal Extractable Value (MEV) markets. MEV refers to the p...
Read analysis →
arXiv: Toward Practical Decentralized Proof-of-Location via Physical Witnessing Zones
A new academic paper, published on arXiv, proposes a framework for decentralized proof-of-location using physical witnessing zones. This is not a regulatory change but a technical proposal that cou...
Read analysis →