Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{207249,
        author = {Abhinav Pandey and Shraddha Bisht and Srashti Pal and Deepak Kumar Pathak and Siddharth Pandey},
        title = {Real-Time AI-Based Multi-Channel Spam Detection System Using Transformer Models},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {136-141},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=207249},
        abstract = {Unsolicited and malicious digital messages have become a serious concern for communication security, particularly because spam, phishing, and fraud attempts now appear across SMS, email, social media, and instant messaging platforms. This paper presents a real-time AI-based spam detection framework using transformer models for contextual message classification. The study compares traditional machine learning classifiers with a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) model using the SMS Spam Collection dataset. Classical models, including Naïve Bayes, Logistic Regression, Random Forest, and Support Vector Machine, produced strong accuracy values between 97% and 98%. However, the fine-tuned BERT model achieved better overall performance, with 99% accuracy, 0.97 spam recall, 0.98 spam F1-score, two false positives, and five false negatives. The confusion matrix further indicates that BERT is more effective in identifying short and context-dependent spam messages because it captures semantic meaning rather than relying only on surface-level word patterns. The findings suggest that transformer-based architectures are suitable for real-time spam filtering, especially in modern multi-channel environments where attackers use personalized and AI-assisted phishing content.},
        keywords = {BERT, Spam Detection, Multi-Channel Security, Transformer Models, F1-Score, Phishing.},
        month = {July},
        }