Measuring Data Protection: A Causal Artificial Intelligence Modeling Approach

Feb 19, 2025

Watch on YouTube

Abstract

This research explores the complex challenge of quantifying data protection, tracing its evolution from ancient ethics to modern cybersecurity. It emphasizes the need for a scientific approach with measurable security properties and a strong theoretical base, highlighting the increasing importance and complexity of safeguarding information in the digital age. Key challenges include evolving threats, the gap between theory and practice, and risk modeling. Societal issues like breaches, surveillance, and regulatory gaps are also addressed.

The research develops a causal model examining security controls, vulnerabilities, and threats to understand data exposure. Built on a literature review and security protection taxonomy, the model uses Directed Acyclic Graphs (DAGs) to visualize causal relationships and is validated through various techniques, including assessing model fit and examining confounding factors. The research also outlines experiments for interventions and counterfactual studies.

Finally, the research identifies future directions, emphasizing standardized data protection metrics and adaptive security systems. It stresses the need for consistent measurement to enable performance comparison and adaptation to evolving threats. The research contributes a systematic approach to studying data protection, from problem identification to model development, validation, and future directions, ultimately aiming to enhance information security.