Plagiarism

What Is Plagiarism?

Plagiarism is the use of another person's text, ideas, processes, results, or algorithms without explicitly acknowledging the original author and source. In academic and professional publishing, it constitutes a serious breach of research integrity, whether the unattributed material is copied verbatim, paraphrased, or presented as an original contribution after minor reformulation. The concept spans multiple disciplines, but it is of particular concern in engineering and scientific research, where the rigorous attribution of methods and results is essential to the reproducibility and credibility of the knowledge record. IEEE defines plagiarism broadly to include the use of any prior work without attribution, regardless of medium or degree of similarity.

Plagiarism is distinct from copyright infringement in that it is fundamentally an ethical violation, while copyright infringement is a legal one; the two can coincide but need not. An author can plagiarize work that is in the public domain, and can infringe copyright while providing full attribution. Both concerns arise in technical publishing, where the same content may simultaneously violate an organization's ethics policy and the intellectual property rights of a copyright holder.

Forms of Plagiarism

Plagiarism is typically classified by the type and degree of copying. Verbatim plagiarism involves reproducing text word-for-word without quotation marks or citation. Paraphrase plagiarism restates the original content in different words while preserving the structure and ideas, without attribution. Mosaic plagiarism interweaves copied phrases among original text in a way that obscures the source. Self-plagiarism, also called duplicate publication, occurs when an author reuses substantial portions of their own previously published work in a new submission without disclosure. In the context of technical conferences and journals, submitting the same paper to multiple venues simultaneously, a practice called dual submission, is treated as a form of self-plagiarism. A PMC comparative analysis of text-based plagiarism detection techniques distinguishes these categories systematically and traces how detection algorithms must be tailored to address each type.

Detection Methods

Automated plagiarism detection relies on text similarity analysis, which compares a submitted document against a corpus of existing publications. Early systems used string matching and fingerprinting approaches, in which fixed-length substrings called n-grams are extracted and compared across documents. Term frequency-inverse document frequency (TF-IDF) weighting and cosine similarity scoring allowed more robust comparison of documents with differing vocabulary. Modern systems increasingly apply natural language processing and neural embedding models to detect paraphrase plagiarism, which defeats character-level matching by rewording rather than copying. An IEEE conference paper on automatic plagiarism detection using natural language processing demonstrates that sentence-level semantic embeddings substantially improve recall over n-gram methods for paraphrased content. IEEE uses the CrossCheck portal, powered by the iThenticate similarity detection service, to screen manuscript submissions against a database of published literature before peer review.

Copyright protection provides the legal framework that underlies many plagiarism sanctions in technical publishing. When a paper is published in an IEEE journal or conference, authors typically transfer copyright to IEEE, which then holds the exclusive right to reproduce and distribute the work. Unauthorized reproduction of substantial portions of a copyrighted paper in a new submission is simultaneously an ethics violation subject to IEEE's internal disciplinary process and a potential copyright infringement. The IEEE guidelines for handling plagiarism complaints outline a graduated investigative process: a complaint is filed, an independent committee reviews similarity evidence, the accused author is given the opportunity to respond, and sanctions are imposed if the complaint is sustained. Sanctions can include retraction of the offending paper, a prohibition on publishing in IEEE venues for a defined period, and placement of the author's name in a prohibited participants list.

Applications

Plagiarism detection and prevention tools are applied across a wide range of contexts, including:

  • Manuscript screening by academic journals and conference program committees
  • University course management systems for checking student submissions
  • Legal proceedings involving copyright disputes over technical documents
  • Corporate research and development to protect trade secrets and verify originality
  • Journalism and media fact-checking to identify sourcing violations