PENGEMBANGAN SISTEM DETEKSI PHISHING PADA WHATSAPP BERBASIS AI DENGAN INTEGRASI BAHASA LOKAL SURABAYA

Phishing is a form of cybercrime that continues to grow, especially through instant messaging platforms such as WhatsApp. Phishing messages often use fake links and persuasive language to deceive victims. This study aims to design and implement a phishing detection system based on Artificial Intelligence (AI) using two fuzzy logic approaches: Fuzzy Mamdani and Fuzzy Takagi-Sugeno-Kang (TSK), integrated into the WhatsApp platform. The system detects phishing based on several parameters, including the number of suspicious keywords, the presence of URLs, and message length, while also considering the local language context such as Indonesian, Javanese, and the Surabaya dialect. The AI model was developed using the FastAPI framework and connected to a WhatsApp bot built with Node.js for real-time communication. Implementation results show that both fuzzy models can accurately detect phishing messages in real time, with confidence levels that are consistent between manual calculations and system outputs. The system also provides automatic alerts to individual users and group chats and stores detection results in a database for further analysis. These findings highlight the effectiveness of AI-based detection systems in enhancing digital communication security, especially when adapted to regional language characteristics.