EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Privacy-Preserving and Trustworthy AI Infrastructures for Digital Commerce

Ahmad Zulkifli Bin Idris, Weiling Tan, Kavitha Rajendran

Journal of Global Trends in Social Science · 31.03.2026 · 3 (2) · 41-53

DOI: 10.70731/1178nq90 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Global Trends in Social Science, online erschienen am 31.03.2026. Verfasst von Ahmad Zulkifli Bin Idris und 2 weiteren Autor:innen. Die Arbeit nennt 57 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Digital commerce is moving from an era of unconstrained data accumulation to one in which privacy regulation, platform restrictions, and public distrust directly shape how artificial intelligence can be designed and deployed. This review synthesizes research on privacy-preserving machine learning, digital advertising measurement, recommender systems, social commerce, creator monetization, platform architecture, and AI governance in order to frame digital commerce as an infrastructure problem rather than a collection of isolated model-level optimizations. The core argument is that privacy-preserving and trustworthy commerce AI requires joint design across four layers: data topology, learning protocol, measurement logic, and governance architecture. At the data layer, the fragmentation of consumer traces across merchants, platforms, creators, and devices makes centralized modeling increasingly costly, risky, and legally fragile. At the learning layer, federated learning, differential privacy, secure aggregation, and related cryptographic tools provide a modular stack for distributed model training and protected analytics, but they also introduce accuracy loss, communication overhead, personalization challenges, and new attack surfaces. At the measurement layer, the decline of third-party identifiers changes attribution, conversion measurement, audience matching, and incentive design; privacy-preserving measurement therefore becomes central to campaign optimization rather than a compliance afterthought. At the governance layer, multi-tenant platform architecture, API standardization, content governance, auditability, and risk management determine whether privacy-enhancing technologies can become operational at scale, especially for small and medium-sized businesses (SMBs) that lack internal AI governance capacity. By integrating evidence from computer science, information systems, marketing, and policy research, the review develops an analytical framework for commerce AI systems that must simultaneously preserve privacy, sustain model utility, support cross-channel decision making, and remain governable. The paper concludes that the next stage of digital commerce AI will be defined less by raw prediction performance than by the ability to institutionalize privacy, transparency, and accountability within commercially viable infrastructures.

Zitieren

APA

Idris, A. Z. B., Tan, W., & Rajendran, K. (2026). Privacy-Preserving and Trustworthy AI Infrastructures for Digital Commerce. Journal of Global Trends in Social Science, 3(2), 41-53. https://doi.org/10.70731/1178nq90

Vancouver

Idris AZB, Tan W, Rajendran K. Privacy-Preserving and Trustworthy AI Infrastructures for Digital Commerce. Journal of Global Trends in Social Science. 2026;3(2):41-53. doi:10.70731/1178nq90

BibTeX
@article{Idris2026,
  author = {Idris, Ahmad Zulkifli Bin and Tan, Weiling and Rajendran, Kavitha},
  title = {Privacy-Preserving and Trustworthy AI Infrastructures for Digital Commerce},
  journal = {Journal of Global Trends in Social Science},
  year = {2026},
  volume = {3},
  number = {2},
  pages = {41--53},
  doi = {10.70731/1178nq90}
}
Kostenlos starten →