EU Regulatory Changes

This publication introduces a novel technical method for detecting Trojan attacks in deep neural networks (DNNs) using spectral regression analysis. While not a regulatory change itself, it represe...

A new research paper published on arXiv proposes an image encryption method using data-identified discrete chaotic maps, which could have implications for data protection and AI safety compliance. ...

This paper, published on arXiv under the AI Safety framework, introduces a novel methodology for managing loss of control risks in advanced AI systems, specifically tailored to national security co...

This publication introduces a new evidence-driven protocol for building trustworthy continuous integration (CI) pipelines, specifically designed to align with the AI Safety framework. The protocol ...

This paper, published on arXiv, proposes a new evidentiary framework for using verifiable provenance and watermarking technologies in generative AI. It specifically addresses how these technical me...