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{193825,
        author = {Prathamesh Mohite and Harsh Pardeshi and Viraj Kamble and Jay Patil and Seema Mishra},
        title = {Authintic: A Three-Layer Hybrid AI System for Plagiarism Detection with Educational Guidance},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {10},
        pages = {2094-2104},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=193825},
        abstract = {The rapid proliferation of Large Language Models (LLMs) such as ChatGPT, Llama, and DeepSeek has created unprecedented challenges in academic integrity. Traditional plagiarism detection tools, which rely primarily on lexical matching, fail to identify semantically paraphrased or AI-generated content. This paper presents Authintic, a full-stack web-based plagiarism detection system that employs a novel three-layer hybrid detection pipeline combining Term Frequency–Inverse Document Frequency (TF-IDF) cosine similarity, FAISS-indexed sentence-level semantic search using Sentence-Transformers, and a fine-tuned BERT binary classifier trained on the PAN 2025 Generative Plagiarism Detection dataset. The system classifies input text at the sentence level into four categories: Direct Match, Paraphrased, AI-Paraphrased, and Original. Additionally, Authintic integrates a Google Gemini-powered educational guidance engine that teaches students how to fix plagiarism rather than rewriting text for them. Evaluated on a balanced PAN25 test set with document-level split integrity, the BERT classifier achieves a Precision of 0.997, Recall of 0.990, F1-Score of 0.9935, and Accuracy of 99.35%, with only 3 false positives and 10 false negatives out of 2,000 test samples.},
        keywords = {Plagiarism Detection, Hybrid Detection, TF-IDF, Sentence Embeddings, FAISS, BERT Fine-tuning, AI-Generated Content Detection, Academic Integrity},
        month = {March},
        }