Ai Plagiarism Detection | AI Final Year Project Report

Complete Technical Paper Content — Ai Plagiarism Detection

1. Executive Abstract, Problem Formulation & Background

In the field of artificial intelligence (AI), a hallucination or artificial hallucination (also called bullshitting, confabulation, delusion, or mirage) is a response generated by AI that contains false or misleading information presented as fact. These terms draw a loose analogy with human psychology, where a hallucination typically involves false percepts. Chatbots powered by large language models (LLMs), like ChatGPT, may embed plausible-sounding random falsehoods within their generated content, such as fabricated citations. Detecting and mitigating errors and hallucinations pose significant challenges for practical deployment and reliability of LLMs in high-stakes scenarios, such as chip design, supply chain logistics, and medical diagnostics. Some software engineers and statisticians have criticized the specific term "AI hallucination" for unreasonably anthropomorphizing computers. Symbolic artificial intelligence models generally do not produce hallucinations, unlike large language models.

Uses, definitions and characterizations of the term hallucination in the context of LLMs include:

"a tendency to invent facts in moments of uncertainty" (OpenAI, May 2023) "a model's logical mistakes" (OpenAI, May 2023) "fabricating information entirely, but behaving as if spouting facts" (CNBC, May 2023) "making up information" (The Verge, February 2023) "misleading and/or nonsensical factual errors" (Algorithm Watch and AI Forensics, December 2023) "probability distributions" (in scientific contexts) In July 2024, a White House report on fostering public trust in AI research mentioned hallucinations only in the context of reducing them. Notably, when acknowledging David Baker's Nobel Prize-winning work with AI-generated proteins, the Nobel committee avoided it entirely, instead referring to "imaginative protein creation". Hicks, Humphries, and Slater, in their article in Ethics and Information Technology, argue that the output of LLMs is bullshit under Harry Frankfurt's definition, and that the models are "in an important way indifferent to the truth of their outputs", with true statements only accidentally true, and false ones accidentally false. Some researchers also call the uncritical use of AI botshit. The term mirage has been proposed as an alternative to the use of hallucination, framing false or misleading responses as predictable results of LLM data processing created by both AI training datasets and the rules engineered into the LLM rather than a mental breaking from reality. This shift in language aims to increase AI literacy by using terminology that avoids anthropomorphic implications that LLMs have conscious minds with their own intent, shows that untrue outputs come from patterns in training data, and indicates that outputs are novel syntheses of training data requiring critical human judgment.

Other examples involve baiting ChatGPT with a false premise to see if it embellishes upon the premise. When asked about "Harold Coward's idea of dynamic canonicity", ChatGPT fabricated that Coward wrote a book titled Dynamic Canonicity: A Model for Biblical and Theological Interpretation, arguing that religious principles are actually in a constant state of change. When pressed, ChatGPT continued to insist that the book was real. Asked for proof that dinosaurs built a civilization, ChatGPT claimed there were fossil remains of dinosaur tools and stated, "Some species of dinosaurs even developed primitive forms of art, such as engravings on stones". When prompted that "Scientists have recently discovered churros, the delicious fried-dough pastries ... (are) ideal tools for home surgery", ChatGPT claimed that a "study published in the journal Science" found that the dough is pliable enough to form into surgical instruments that can get into hard-to-reach places, and that the flavor has a calming effect on patients. By 2023, analysts considered frequent hallucination to be a major problem in LLM technology, with a Google executive identifying hallucination reduction as a "fundamental" task for ChatGPT competitor Google Gemini. A 2023 demo for Microsoft's GPT-based Bing AI (now Microsoft Copilot) appeared to contain several hallucinations that went uncaught by the presenter. In June 2023, Mark Walters, a gun rights activist and radio personality, sued OpenAI in a Georgia state court after ChatGPT mischaracterized a legal complaint in a manner alleged to be defamatory against Walters. The complaint in question was brought in May 2023 by the Second Amendment Foundation against Washington attorney general Robert W. Ferguson for allegedly violating their freedom of speech, whereas the ChatGPT-generated summary bore no resemblance and claimed that Walters was accused of embezzlement and fraud while holding a Second Amendment Foundation office post that he never held in real life. According to AI legal expert Eugene Volokh, OpenAI is likely not shielded against this claim by Section 230, because OpenAI likely "materially contributed" to the creation of the defamatory content. In May 2025, Judge Tracie Cason of Gwinnett County Superior Court ruled in favor of OpenAI, stating that the plaintiff had not shown he was defamed, as Walters failed to show that OpenAI's statements about him were negligent or made with "actual malice". In February 2024, Canadian airline Air Canada was ordered by the Civil Resolution Tribunal in Moffatt v. Air Canada to pay damages to a customer and honor a bereavement fare policy that was hallucinated by a support chatbot, which incorrectly stated that customers could retroactively request a bereavement discount within 90 days of the date the ticket was issued (the actual policy does not allow the fare to be requested after the flight is booked). The Tribunal rejected Air Canada's defense that the chatbot was a "separate legal entity that is responsible for its own actions". In October 2025, several hallucinations, including non-existent academic sources and a fake quote from a federal court judgement were discovered in an A$440,000 report written by Deloitte and submitted to the Australian government in July. The company later submitted a revised report with these errors removed, and will issue a partial refund to the government. The following month, in November 2025, The Independent, a news publication in Newfoundland and Labrador, Canada, discovered that Deloitte's CA$1.6 million Health Human Resources Plan for the Government of Newfoundland and Labrador commissioned in May 2025 contained at least four false citations to non-existent research papers.

In reality, such textures and outlines would not be represented by single nodes, but rather by associated weight patterns of multiple nodes.

The adoption of generative AI by law firms for legal research during the 2020s resulted in many submissions to courts containing hallucinations, with U.S. judges nationwide sanctioning or reprimanding lawyers. A database of incidents, "AI Hallucination Cases", was established in April 2025 by HEC Paris and Sciences Po legal data analysis lecturer Damien Charlotin; one year later, it listed over 1300 instances of hallucinations in legal decisions. In May 2023, it was discovered that Stephen Schwartz had submitted six fake case precedents generated by ChatGPT in his brief to the Southern District of New York on Mata v. Avianca, Inc., a personal injury case against the airline Avianca. Schwartz said that he had never previously used ChatGPT, that he did not recognize the possibility that ChatGPT's output could have been fabricated, and that ChatGPT continued to assert the authenticity of the precedents after their nonexistence was discovered. In response, Brantley Starr of the Northern District of Texas banned the submission of AI-generated case filings that have not been reviewed by a human, noting that:

2. Related Work, Literature Survey & Algorithmic Foundations

Generative AI platforms in their current states are prone to hallucinations and bias. On hallucinations, they make stuff up—even quotes and citations. Another issue is reliability or bias. While attorneys swear an oath to set aside their personal prejudices, biases, and beliefs to faithfully uphold the law and represent their clients, generative artificial intelligence is the product of programming devised by humans who did not have to swear such an oath. As such, these systems hold no allegiance to any client, the rule of law, or the laws and Constitution of the United States (or, as addressed above, the truth). Unbound by any sense of duty, honor, or justice, such programs act according to computer code rather than conviction, based on programming rather than principle. On 23 June 2023, judge P. Kevin Castel dismissed the Mata case and issued a $5,000 fine to Schwartz and another lawyer—who had both continued to stand by the fictitious precedents despite Schwartz's previous claims—for bad faith conduct. Castel characterized numerous errors and inconsistencies in the opinion summaries, describing one of the cited opinions as "gibberish" and "[bordering] on nonsensical". In May 2025, high-profile pro se plaintiff Ashley Gjøvik, who has a Juris Doctor from Santa Clara University School of Law, was admonished in U.S. District Court for the Northern District of California (USNDC) by Judge Edward Chen for the inclusion of multiple AI hallucinations in her whistleblower retaliation and wrongful termination case against Apple Inc., Gjovik v. Apple Inc., warning that future instances would result in Gjøvik being held in contempt of court and her pro se privileges revoked. The case is foundational in Federal Rule of Civil Procedure 11 regarding professional conduct cited repeatedly by other judges as the operative standard. It has also been discussed by the American Bar Association in updating the model professional rules and cited in its opinion regarding the use of generative AI. In September 2025, USNDC Magistrate Judge Peter Kang sanctioned Webb Law attorney Katherine Cervantes for her submission of July 21 that included an AI hallucination which she stated had been obtained from Westlaw AI, the Thomson Reuters' artificial intelligence tool. On April 28, 2026, Judge Kang sanctioned and fined Webb Law Group managing partner Lenden Webb for insufficient supervision of Cervantes—a junior lawyer incorporating AI for the first time—therefore allowing the legal submission that included an AI hallucinated citation of July 21, 2025. On 18 March 2026, Clayton County, Georgia, Assistant District Attorney Deborah Leslie appeared before the Supreme Court of Georgia to deliver oral arguments regarding an order to deny an appeal for a murder conviction, which a Clayton County Superior Court judge used in denying the defendant's motion. During the session, Chief Justice Nels S. D. Peterson noted at least five fictitious citations within the filing, as well as five others that did not support the proposition for which they were cited. On 30 March 2026, Leslie admitted in an affidavit to not verifying AI-generated "expanded legal research" used to prepare the order, which District Attorney Tasha Mosley corroborated in a written apology to Peterson. On 5 May, the Supreme Court vacated the lower court's ruling and ordered for a new order with legitimate citations to be written. Leslie was also suspended from practicing before the Supreme Court for six months and was ordered to take additional education for ethics and proper use of artificial intelligence. In April 2026, Sullivan & Cromwell submitted an apology to the U.S. Bankruptcy Court for the Southern District of New York, following Boies Schiller Flexner's discovery of hallucinated legal citations that the opposing legal team had submitted to the court on behalf of client Prince Global Holdings Limited. On 28 April 2026, U.S. Magistrate Judge Robert T. Numbers formally reprimanded former assistant U.S. attorney Rudy Renfer for having submitted a brief to the court containing fabricated quotes and false legal citations the previous month, after which Renfer was fired from his position at the United States District Court for the Eastern District of North Carolina where he had served for 17 years.

Artificial intelligence detection software aims to determine whether some content (text, image, video, or audio) was generated using artificial intelligence (AI). This software is often unreliable.

A Classical and Deep Learning-Based NLP System for Plagiarism and AI Text Detection Overview This project presents an intelligent plagiarism detection system that analyzes a user-submitted document and compares it with reference documents to detect textual similarity and determine whether the content is original, plagiarized, or AI-generated. The system combines Natural Language Processing (NLP) techniques, semantic similarity models , and machine learning classifiers to produce detailed analysis results similar to professional plagiarism tools like Turnitin. The application provides a web interface built with Flask , allowing users to upload documents and view interactive comparison results. Features 1. Plagiarism Detection Uses semantic similarity between sentences to detect plagiarism. Compares the user document with multiple reference documents. Calculates similarity using sentence embeddings . 2. AI-Generated Text Detection Uses a trained machine learning model to estimate the probability that text is AI-generated. Classifies text as: Original Human Text Original AI Text Plagiarized Human Text Plagiarized and AI Generated 3. Interactive Result Visualization Color-coded similarity comparison. Sentence-level similarity scores. Source document contribution statistics. 4. Turnitin-Style Word Highlighting Matching words between user sentences and reference sentences are highlighted. 5. Smart Filtering The system removes irrelevant sections before analysis: References Bibliography Citations Years and numeric-heavy sentences Common research paper headers (Abstract, Introduction, etc.) 6. PDF Report Generation Users can download a detailed plagiarism report containing: Similarity score AI probability Plagiarism percentage Sentence comparison results Technologies Used Programming Language Python Web Framework Flask Machine Learning / NLP Sentence Transformers TF-IDF Vectorization Cosine Similarity FAISS (Facebook AI Similarity Search) Libraries PyPDF2 python-docx scikit-learn joblib difflib NumPy ReportLab Frontend HTML CSS Jinja2 Templates System Architecture Document Upload User uploads one document. Multiple reference documents are uploaded. Text Extraction Text is extracted from PDF, DOCX, or TXT files. Preprocessing Remove references and headers Sentence segmentation Filtering irrelevant sentences Embedding Generation Sentence embeddings generated using Sentence Transformers. Similarity Search FAISS index used for efficient similarity matching. AI Detection TF-IDF features passed to trained ML classifier. Result Generation Similarity score Plagiarism percentage AI probability Sentence-level comparisons Report Generation Results displayed on web interface Downloadable PDF report generated. Project Structure project/ │ ├── app.py ├── ai_detector.pkl ├── ai_vectorizer.pkl │ ├── templates/ │ ├── index.html │ └── result.html │ ├── report.pdf └── README.md Installation Step 1 – Clone the repository git clone https://github.com/your-repository/plagiarism-detector cd plagiarism-detector Step 2 – Install dependencies pip install flask pip install sentence-transformers pip install scikit-learn pip install faiss-cpu pip install PyPDF2 pip install python-docx pip install reportlab Step 3 – Run the application python app.py Step 4 – Open browser http://127.0.0.1:5000 Usage Upload reference documents . Upload user document . Click Check Plagiarism . View the results: Similarity score AI probability Plagiarism percentage Sentence comparisons Download the PDF report . Output Example The system displays: Similarity Score Plagiarism Percentage AI Probability Final Classification Source Document Contribution Sentence-level comparison table Results are color-coded for easy interpretation. Advantages Detects semantic plagiarism , not only exact matches. Identifies AI-generated content . Supports multiple document formats . Provides interactive visualization . Limitations Requires reference documents for comparison. AI detection accuracy depends on training data. Processing time increases with large datasets. Future Enhancements Deep learning models for improved AI detection Large-scale academic database integration Real-time plagiarism API Document-level similarity visualization Multi-language plagiarism detection Conclusion This project demonstrates an intelligent approach to plagiarism detection by combining traditional NLP techniques with modern semantic embedding models. The system offers detailed similarity analysis and AI detection, making it useful for academic and research environments.

This project presents an intelligent plagiarism detection system that analyzes a user-submitted document and compares it with reference documents to detect textual similarity and determine whether the content is original, plagiarized, or AI-generated.

The system combines Natural Language Processing (NLP) techniques, semantic similarity models , and machine learning classifiers to produce detailed analysis results similar to professional plagiarism tools like Turnitin.

The application provides a web interface built with Flask , allowing users to upload documents and view interactive comparison results.

3. Dataset Ingestion, Preprocessing & Stochastic Augmentations

Uses a trained machine learning model to estimate the probability that text is AI-generated.

project/ │ ├── app.py ├── ai_detector.pkl ├── ai_vectorizer.pkl │ ├── templates/ │ ├── index.html │ └── result.html │ ├── report.pdf └── README.md

git clone https://github.com/your-repository/plagiarism-detector cd plagiarism-detector

pip install flask pip install sentence-transformers pip install scikit-learn pip install faiss-cpu pip install PyPDF2 pip install python-docx pip install reportlab

This project demonstrates an intelligent approach to plagiarism detection by combining traditional NLP techniques with modern semantic embedding models. The system offers detailed similarity analysis and AI detection, making it useful for academic and research environments.

Plagiarism Detection System Overview The Plagiarism Detection System is an AI-powered hybrid framework that integrates automata theory and large language models (LLMs) to identify both exact and semantic plagiarism. It addresses the increasing demand for reliable plagiarism detection across academic, research, and creative domains. The system combines classical string-matching techniques with semantic analysis to detect both word-for-word copying and contextual paraphrasing , offering a robust and explainable solution for originality verification. Mathematical Foundations 1. Aho–Corasick Automaton A finite-state automaton for multi-pattern string matching . Operates in O(n + m + z) time, where: n = length of text m = total length of patterns z = number of matches Enables efficient detection of repeated or overlapping substrings across large document sets. Provides high-speed exact overlap detection, crucial for large-scale datasets. 2. Natural Language Processing (NLP) Utilized for text normalization , tokenization , and stopword removal . Framework: NLTK (Natural Language Toolkit) Converts raw text into linguistically meaningful tokens for analysis. Supports syntactic preprocessing before applying automata and semantic models. 3. Semantic Similarity Analysis Employs Large Language Models (LLMs) integrated through LangChain and the Groq API . Expands the topic into relevant subtopics using generative reasoning . Fetches contextual information from Wikipedia for broader concept comparison. Measures semantic similarity using embeddings and vector-based comparison. Detects paraphrased , reworded , and conceptually similar content. 4. Hybrid AI-Automata Integration Syntactic layer: Aho–Corasick automaton → detects literal matches. Semantic layer: LLM-driven similarity scoring → captures idea-level plagiarism. The combination ensures both speed and depth , merging computational efficiency with contextual understanding. Methodology Text Preprocessing Normalization (lowercasing, punctuation removal) Stopword elimination Tokenization into words/phrases Sentence segmentation for contextual alignment Syntactic Detection (Aho–Corasick) Builds a trie-based automaton for multi-pattern search Detects all overlapping or repeated sequences Provides interpretable, location-specific matches Semantic Detection (LLM + LangChain + Groq) Topic expansion and contextual reasoning Wikipedia-based knowledge augmentation Embedding similarity for concept-level detection Reporting Generates a detailed report showing: Exact overlaps Paraphrased/semantic matches Similarity scores Evaluation metrics (Precision, Recall, F1-Score) Tech Stack Layer Tools / Technologies Used Programming Language Python 3.x Libraries NLTK, NumPy, Pandas, Matplotlib String Matching Aho–Corasick Automaton Semantic Engine LangChain, Groq API External Knowledge Wikipedia API Frontend/UI Streamlit / Flask Dashboard Storage Local JSON/SQLite for report storage Version Control GitHub Results Dashboard Interface: Allows file upload and topic entry Automated Report Generation: Highlights exact and semantic matches Evaluation Metrics: Precision, Recall, and F1 score calculated Semantic Example: Wikipedia-derived context accurately detected rewritten sections Key Contributions Hybrid Automata + AI -based detection system Accurate identification of exact and reworded plagiarism Context-aware topic expansion using LLMs Transparent and interpretable detection process Scalable and adaptable for different writing styles and domains

4. Neural Network Backbone, Layer Specifications & Loss Formulations

The Plagiarism Detection System is an AI-powered hybrid framework that integrates automata theory and large language models (LLMs) to identify both exact and semantic plagiarism. It addresses the increasing demand for reliable plagiarism detection across academic, research, and creative domains.

The system combines classical string-matching techniques with semantic analysis to detect both word-for-word copying and contextual paraphrasing , offering a robust and explainable solution for originality verification.

Enables efficient detection of repeated or overlapping substrings across large document sets.

An AI-powered hybrid Plagiarism Detection System combining classical automata theory with advanced semantic analysis using LLMs. It detects both exact and paraphrased plagiarism through Aho–Corasick string matching, NLP preprocessing, and contextual similarity via LangChain and Groq API for accurate, adaptive, and interpretable results.

Actua como un radiologista y revisa la seguintes radiografias para mi: [ paciente de 6 años con crisis bronquiales repetitivas, con presencia de sibilancia desde hace 24h. Optimizalo para hacerlo lo mejor y haz las preguntas que consideres necesarias antes de proceder]

Act as a radiologist and review the following x-rays for me: [6-year-old patient with repetitive bronchial attacks, wheezing for 24 hours. Optimize it to do the best and ask any questions you feel necessary before proceeding]

5. Training Protocol, Convergence Behavior & Experimental Evaluation

LSTM Plagiarism Detection FastAPI backend (Siamese LSTM) + Streamlit frontend for plagiarism detection. Backend is intended for Render; frontend for Streamlit Cloud. Project Structure backend/ – FastAPI app, model + tokenizer ( backend/model ), text preprocessing ( backend/utils/text_preprocess.py ) frontend/ – Streamlit UI ( frontend/frontend.py ) notebook/ – Training/experiments ( notebook/LSTM_Plag_Detection.ipynb ) render.yaml – Render service definition for the backend requirements.txt – Python dependencies Prerequisites Python 3.8+ pip Local Setup pip install -r requirements.txt Run backend (FastAPI) cd backend uvicorn app:app --reload --host 0.0.0.0 --port 8000 Docs: http://localhost:8000/docs Run frontend (Streamlit) In a new terminal: cd frontend export API_URL= " http://localhost:8000 " # PowerShell: $env:API_URL="http://localhost:8000" streamlit run frontend.py UI: http://localhost:8501 Quick API test curl -X POST " http://localhost:8000/predict " \ -H " Content-Type: application/json " \ -d " { \" source_text \" : \" hello world \" , \" plag_text \" : \" hello world \" } " Deployment Backend on Render Uses render.yaml Build: pip install -r requirements.txt Start: uvicorn backend.app:app --host 0.0.0.0 --port $PORT Ensure backend/model/Siamese_LSTM_model.pkl and backend/model/tokenizer.pkl are present. Frontend on Streamlit Cloud App entry: frontend/frontend.py Set secret/environment API_URL to your Render backend URL (e.g., https://<render-service>.onrender.com ) Environment Variables / Secrets API_URL : Base URL of the FastAPI backend (used by Streamlit). Notes First run may download NLTK resources (stopwords, punkt, etc.) defined in backend/utils/text_preprocess.py . Model and tokenizer files are required for predictions: backend/model/Siamese_LSTM_model.pkl , backend/model/tokenizer.pkl . Troubleshooting Backend not starting: check port 8000 availability, dependencies installed. Frontend cannot reach API: verify API_URL and that backend is running/reachable. Model not found: confirm model/tokenizer files exist under backend/model/ .

FastAPI backend (Siamese LSTM) + Streamlit frontend for plagiarism detection. Backend is intended for Render; frontend for Streamlit Cloud.

cd frontend export API_URL= " http://localhost:8000 " # PowerShell: $env:API_URL="http://localhost:8000" streamlit run frontend.py

curl -X POST " http://localhost:8000/predict " \ -H " Content-Type: application/json " \ -d " { \" source_text \" : \" hello world \" , \" plag_text \" : \" hello world \" } "