Intelligent Multi-Modal Spam Detection System Using Machine Learning Techniques for AI-Generated Content
International Journal of Advanced Research in Science Communication and Technology · 24.06.2026 · 58
DOI: 10.48175/ijarsct-37208 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 24.06.2026. Verfasst von Harsh Rakesh Walmiki. Die Arbeit nennt 30 Literaturangaben.
Abstract (Originalsprache)
Although spam is not a new problem, the emergence of generative AI tools has made the problem more hazardous than ever before. Spam messages written using large language models are extremely convincing since they sound like human messages. In this paper, we present our spam detection system which uses classical NLP text processing techniques and supervised machine learning with the emerging threats of spam in mind. The proposed solution consists of five text preprocessing steps for the SMS messages—lowercasing, noise removal using regular expressions, stopwords removal, and Porter stemming—to convert SMS messages to feature vectors of TF-IDF using up to 5,000 features. Four classifiers have been tested on the UCI SMS Spam Collection dataset which contains 5,572 SMS messages—the Multinomial Naïve Bayes, Logistic Regression, Linear SVM, and Random Forest. The Linear SVM worked best, with an accuracy of 98.2%, 97.3% precision, 90.6% recall, and an F1 score of 93.8%. The part which we find particularly intriguing, and which occupies much of the remainder of this paper, is how this sort of pipeline can still hold its own when it comes to defending against spam generated by AI. This is because stemming reduces the diversity of the vocabulary exploited by LLMs, IDF weighting automatically discourages the common tokens favored by language models, and the geometry of margins in SVM gives some leeway in case of perturbation of feature space. We have also developed a web application using Flask for real-time predictions.Index Terms— Spam Filtering, Text Classification, NLP, TF-IDF, SVM, Naïve Bayes, AI-Generated Content, Adversarial Robustness
Zitieren
APA
Walmiki, H. R. (2026). Intelligent Multi-Modal Spam Detection System Using Machine Learning Techniques for AI-Generated Content. International Journal of Advanced Research in Science Communication and Technology, 58. https://doi.org/10.48175/ijarsct-37208
Vancouver
Walmiki HR. Intelligent Multi-Modal Spam Detection System Using Machine Learning Techniques for AI-Generated Content. International Journal of Advanced Research in Science Communication and Technology. 2026:58. doi:10.48175/ijarsct-37208
BibTeX
@article{Walmiki2026,
author = {Walmiki, Harsh Rakesh},
title = {Intelligent Multi-Modal Spam Detection System Using Machine Learning Techniques for AI-Generated Content},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {58},
doi = {10.48175/ijarsct-37208}
}
EUVIMED