Engineering Secure AI/ML Systems: Implementing Cloud-Based Differential Privacy Strategies for Enhanced Security

EasyChair Preprint 15013

Abstract

As artificial intelligence (AI) and machine learning (ML) technologies become integral to various industries, ensuring the security and privacy of sensitive data is paramount. This article explores the implementation of cloud-based differential privacy strategies as a robust framework for engineering secure AI/ML systems. By leveraging differential privacy, organizations can effectively protect individual data points while still enabling meaningful data analysis and model training. The discussion highlights key principles of differential privacy, its integration into cloud environments, and practical applications across sectors such as healthcare, finance, and social media. Furthermore, the article addresses challenges associated with deploying these strategies, including computational overhead and the trade-offs between privacy and utility. Through a series of case studies, we illustrate successful implementations that demonstrate the effectiveness of cloud-based differential privacy in safeguarding user data while maintaining the performance of AI/ML systems. This comprehensive examination aims to provide industry stakeholders with actionable insights and best practices for enhancing data security in an increasingly interconnected digital landscape.

Keyphrases: Enhancing, Security, Stakeholders, actionable, data, insights, user

@booklet{EasyChair:15013,
  author    = {Kayode Sheriffdeen},
  title     = {Engineering Secure AI/ML Systems: Implementing Cloud-Based Differential Privacy Strategies for Enhanced Security},
  howpublished = {EasyChair Preprint 15013},
  year      = {EasyChair, 2024}}