Secure Communications Platform With Aiml - Based Threat Detection

• Author(s):

  Adithya Varshan A | Ashvanthgopal Baskaran | Ranjith V | Dr. V. Ravindra Krishna Chandar

• Keywords:

  Spam Call Detection, Machine Learning, Naive Bayes, TF-IDF, Multilingual NLP, Speech-to-text, AI Screener, Cybersecurity, Indian Languages

• Abstract:

  The Increasing Prevalence Of Spam And Scam Calls Poses A Significant Security Threat, Particularly In Multilingual Regions Such As India. This Paper Presents ShieldCall AI, An AI-driven Spam Call Detection System That Integrates Machine Learning, Natural Language Processing (NLP), And Real-time Audio Processing To Identify Fraudulent Calls Across Eight Major Indian Languages. The System Employs A Multinomial Naive Bayes Classifier With TF-IDF Feature Extraction To Classify Call Transcripts As Spam Or Legitimate. Live Audio Is Transcribed Via Google Speech Recognition, While Language Detection And Translation Modules Support Hindi, Tamil, Telugu, Kannada, Malayalam, Bengali, Marathi, And Gujarati By Translating Regional Language Transcripts To English Prior To Classification. The Backend Is Implemented Using Python And Flask, And The Frontend Is A Mobile-first Web Application Deployed On Render. A Distinctive Feature Is The Live AI Screener, An Autonomous Conversational Agent That Interacts With Suspected Spam Callers Using Voice Synthesis, Collects Evidence, And Dynamically Escalates Risk Scores. Testing Across All Supported Languages Achieved High Spam Detection Accuracy With Low False-positive Rates, Demonstrating The Practical Effectiveness Of The System For Real-world Personal And Organizational Security Applications

---

Other Details

• Paper id:

  IJSARTV12I5105437

• Published in:

  Volume: 12 Issue: 5 May 2026

• Publication Date:

  2026-05-22

Download Article