Leveraging Homomorphic Encryption to Enhance Efficiency and Data Protection in Online Retail Based in Kenya
Mukolwe N Lucy, Philemon Kittur and Francis Musembi
Department of Mathematics and Computer science, University of Eldoret, Kenya.
Email: lucymukolwe@gmail.com
Abstract: The rapid growth of online retail businesses has increased reliance on artificial intelligence (AI)-driven systems for personalization, fraud detection, and operational efficiency in customer interaction. These algorithmic systems require access to vast amounts of sensitive customer data, positioning data as both a driver of commercial value and a source of digital vulnerability. In many cases, this data is processed in plain text, creating significant security risks especially in dynamic and cloud-based retail environments. This exposes critical shortcomings in traditional encryption methods and raises fundamental concerns about digital rights, trust, and data governance in the algorithmic age. This study explores the integration of Partial Homomorphic Encryption (PHE) as a means of reinforcing data privacy without sacrificing functionality. PHE allows for computations on encrypted data, eliminating the need for decryption during processing and thereby preserving customer confidentiality. The research focuses on current data security practices in online retail businesses, with particular attention to Kenyan-based enterprises, identifying key vulnerabilities in the handling of sensitive customer information. It further examines the limitations of existing traditional encryption approaches and presents a theoretical application of PHE within AI models deployed by online retailers. Through a literature review, the study assesses the performance, feasibility, and implications of adopting Homomorphic Encryption in cloud-based retail systems. The findings reveal that although PHE introduces computational overhead, this research argues that embracing PHE can help redistribute power over personal data, reinforce consumer trust, and uphold digital rights in the increasingly algorithmic architecture of online retail systems.
