MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202641079479 A) filed by Kosegi Amarr Kannthh Reddi; Benda Hari Kumar; Kampati Naveen Kumar; J Bangaru Siddhartha; Smitha Sharath Shankar; M. Dhanraj; Kusumanchi Bhuvaneshwar; Dr. Anoop Kumar Tiwari; Dr. Pawan Singh; S. D. Anushna; and M. Ishwariya on June 28, 2026, for "device And Method Ai-Powered Models For Predicting Fake News Virality On Social Media Networks".
Inventors include Kosegi Amarr Kannthh Reddi; Benda Hari Kumar; Kampati Naveen Kumar; J Bangaru Siddhartha; Smitha Sharath Shankar; M. Dhanraj; Kusumanchi Bhuvaneshwar; Dr. Anoop Kumar Tiwari; Dr. Pawan Singh; S. D. Anushna; and M. Ishwariya.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: ABSTRACT OF THE INVENTION: The present invention provides a device and method for AI-powered prediction of fake news virality on social media networks. The invention addresses the critical challenge of identifying which misinformation content is likely to spread rapidly and widely, enabling proactive intervention before widespread dissemination occurs. The device comprises a data ingestion module for collecting social media content and metadata, a multilingual and code-mixed preprocessing module that processes content in multiple languages with support for both Devanagari and Roman scripts, and a multimodal feature extraction engine that analyzes textual, visual, and metadata content. A propagation network construction module creates hierarchical propagation networks comprising macro-level networks of news nodes and social media post nodes, and micro-level networks of reply nodes. The system analyzes structural features including cascade depth, breadth, and structural virality, along with temporal features such as propagation speed and timing patterns. A predictive analytics engine utilizes a hybrid neural network architecture with dual encoding streams,a text encoder generating contextual embeddings and a structured feature stream processing source credibility scores, social engagement metrics, psycholinguistic indicators, sensationalism markers, and sentiment features. The engine generates virality risk scores predicting content's propagation potential. An incremental learning module enables real-time adaptation to evolving misinformation patterns through continuous model updating without full retraining. An explanation generation module provides human-readable justifications for predictions using SHAP analysis, saliency maps, or rule-based reasoning, ensuring transparency and trust for non-technical stakeholders. The device further includes a dashboard interface with role-specific views for researchers, journalists, policymakers, and social media platform moderators. The dashboard presents virality predictions, risk assessments, propagation visualizations, and actionable insights for proactive intervention strategies. The present invention provides a comprehensive technological solution for addressing the societal challenge of viral misinformation. By combining multilingual and multimodal analysis, hierarchical propagation network modeling, predictive virality scoring, real-time adaptive learning, and explainable AI, the invention enables early detection and intervention in the spread of fake news, supporting more effective content moderation, fact- checking prioritization, and public awareness campaigns. ________________________________________
Disclaimer: Curated by HT Syndication.
