Evaluation of cheating detection methods in academic writings

Purpose - This paper aims to focus on plagiarism and the consequences of anti-plagiarism services such as Turnitin.com, iThenticate, and PlagiarismDetect.com in detecting the most recent cheatings in academic and other writings. Design/methodology/approach - The most important approach is plagiarism prevention and finding proper solutions for detecting more complex kinds of plagiarism through natural language processing and artificial intelligence self-learning techniques. Findings - The research shows that most of the anti-plagiarism services can be cracked through different methods and artificial intelligence techniques can help to improve the performance of the detection procedure. Research limitations/implications - Accessing entire data and plagiarism algorithms is not possible completely, so comparing is just based on the outputs from detection services. They may produce different results on the same inputs. Practical implications - Academic papers and web pages are increasing over time, and it is very difficult to capture and compare documents with all available data on the network in an up to date manner. Originality/value - As many students and researchers use the plagiarism techniques (e.g. PDF locking, ghost-writers, dot replacement, online translators, previous works, fake bibliography) to cheat in academic writing, this paper is intended to prevent plagiarism and find suitable solutions for detecting more complex kinds of plagiarism. This should also be of grave concern to teachers and librarians to provide up to date/standard anti-plagiarism services. The paper proposes some new solutions to overcome these problems and to create more resilient and intelligent future systems.