Year: 2026
Type: article
Abstract: A growing problem across the globe is the digital age-old dictating that modern technological developments have enabled: the use of advanced AI techniques like deepfake technology, large language models, real-time speech synthesis, and auto-phishing systems is heavily used to defraud private individuals, businesses, and public institutions in larger proportions than ever before. Traditional cybersecurity tools focused on preventing threats at the network perimeter or through rules are structurally weak to fight this attack. The author contends that all cities no matter their size, digital readiness, or economic condition should have their own AI Scam Shield, an intelligent cross-sectorial, real-time threat recognition, behavioural analysis, and community protection system designed on a city scale to thwart AI scams before they can harm citizens. A systematic synthesis of literature covering the field of fraud detection using AI, smart city security architecture, protection of the elderly population, phishing protection using Blockchains, and innovative regulations, revealed that modern protections are not adequate. The study outlines a complete framework of how to implement municipal AI Scam Shield. The results show integrated AI Scam Shield systems to be more accurate at detecting phishing frauds with an accuracy of 96.4%, have less delay in real-time fraud alerts of more than 99% as compared to machine-learning-driven systems, and were able to intercept elderly-targeted phishing attempts almost double, at 85.4% versus 40.9%. Finally, the article details the governance framework and policy roadmap for the implementation of AI Scam Shield at the City level, highlighting governance principles of equity, transparency, and civil rights protection.
Author: Rohit Rajdev
Language: English
Topic: Spam and Phishing Detection
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7162079390
FWCI:
Cites:
Cited by:
Related to: