AI-Based Real -Time Screen Reading and Alert System

Abstract

The rise in cyber threats like phishing websites, malicious URLs, harmful messages, false login pages, and scam advertisements has increased the probability of identity theft and frauds to a great extent. Existing security systems mostly employ browser-based security and signature based detection methods which have lower accuracy levels when it comes to detecting new forms of cyberattacks and displaying any malicious content on the computer screen. In this paper, we present the idea of an AI-Based Real Time Screen Reading & Alerting System that continuously monitors the user screen and checks for any phishing attack, suspicious URLs, harmful messages, and offensive text using Optical Character Recognition (OCR) and machine learning. The suggested system regularly takes screenshots, retrieves the text from the screenshot using the Tesseract OCR library, and then analyzes the text using NLP. It includes a phishing URL detector which checks the suspicious links found in the screenshots against malicious URL datasets and a keyword analysis module to find any harmful or suspicious content. If any threat is detected, then an alert is issued to the user to proactive cybersecurity by providing real-time protection against online threats.