AI-Augmented Cloud Security Posture Management for Securing Enterprise AI Workloads
International Journal of Scientific Research in Computer Science, Engineering and Information Technology · 30.06.2024 · 10 (3) · 1171-1184
DOI: 10.32628/cseit25113585 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Scientific Research in Computer Science, Engineering and Information Technology, online erschienen am 30.06.2024. Verfasst von Lakshmi Kiran Meesala. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Cloud Security Posture Management (CSPM) has historically targeted infrastructure misconfigurations - exposed storage buckets, overprivileged IAM roles, and open network ports. The proliferation of AI workloads across enterprise clouds introduces fundamentally novel attack vectors: GPU resource hijacking, model training data exfiltration, shadow AI deployments, prompt injection surfaces, and insecure ML pipeline configurations. Existing CSPM frameworks lack semantics for AI asset discovery, AI-specific policy enforcement, and behavioral anomaly detection across ephemeral GPU compute clusters. This paper proposes AI-Aware CSPM (AA-CSPM), an architectural extension that integrates ML-based risk scoring, AI-specific misconfiguration detection rules, data lineage monitoring, and convergent posture management across CSPM, Data Security Posture Management (DSPM), and Cloud Infrastructure Entitlement Management (CIEM). We formalize a threat taxonomy of twelve AI-era attack classes, implement detection pipelines across three major cloud providers (AWS SageMaker, Google Vertex AI, Azure ML), and benchmark AA-CSPM against three baseline CSPM platforms. AA-CSPM achieves a mean detection accuracy of 94.3%, a false-positive reduction of 38.7%, and reduces mean time-to-detect (MTTD) for AI-specific misconfigurations by 61.4% over legacy baselines. These results demonstrate that AI workloads constitute a distinct and underserved risk domain that demands dedicated posture management tooling.
Zitieren
APA
Meesala, L. K. (2024). AI-Augmented Cloud Security Posture Management for Securing Enterprise AI Workloads. International Journal of Scientific Research in Computer Science, Engineering and Information Technology, 10(3), 1171-1184. https://doi.org/10.32628/cseit25113585
Vancouver
Meesala LK. AI-Augmented Cloud Security Posture Management for Securing Enterprise AI Workloads. International Journal of Scientific Research in Computer Science, Engineering and Information Technology. 2024;10(3):1171-1184. doi:10.32628/cseit25113585
BibTeX
@article{Meesala2024,
author = {Meesala, Lakshmi Kiran},
title = {AI-Augmented Cloud Security Posture Management for Securing Enterprise AI Workloads},
journal = {International Journal of Scientific Research in Computer Science, Engineering and Information Technology},
year = {2024},
volume = {10},
number = {3},
pages = {1171--1184},
doi = {10.32628/cseit25113585}
}
EUVIMED