The Anti-Plagiarism text verification system has learned to recognize materials written by neural networks.  The company's developers managed to achieve 98% accuracy in determining text written by artificial intelligence (AI), which is 35% higher than the previous indicators. The improved program has been tested and trained on the most popular neural networks, such as deepseekV3 and GPT-4o.

The detection algorithm is an in-house development of the Antiplagiat company, created on the basis of many years of research on natural language processing ( NLP) and is included in the standard license for text verification for the detection of borrowings. The improved algorithm of machine-generated texts has been trained and tested on a large data set, which minimizes the risk of false positives.

The new Anti-Plagiarism algorithm is designed specifically for working with academic and scientific texts. He is trained on a unique selection from the collections collected by the company over 20 years of cooperation with universities. This makes it accurate in identifying AI content in diplomas and scientific publications, where originality and authenticity are critically important.

The new, free Chinese deepseekV3 neural network has become the most popular AI-based program in Russia since its launch in January 2025 , increasing its mobile traffic 5 times. At the same time, the American OpenAI introduced its new neuromodel GPT-4o.

A recent study by the company "I am a professional" says that 85% of Russian students regularly use neural networks. Thus, 77% of students use AI to search for information , 43% use it to write research papers, essays, term papers and theses , and 24% of students use it to create presentations. Previously, there have been cases when students of Russian universities defended a diploma written by a neural network. At the moment, almost all national and federal universities in Russia use a text uniqueness recognition system.

92% of Priority 2030 universities use "Anti-plagiarism". Among them are Lomonosov Moscow State University, Saint Petersburg State University, HSE, Bauman Moscow State Technical University, Plekhanov Russian University of Economics and many other leading universities in the country.

"We hope that this update will encourage the authors to focus on developing their own writing and analysis skills, which, in turn, will help them move within the framework of academic ethics and ensure high standards of Russian education. We are confident that in the long term, the new detection algorithm will bring much more benefits than fast but unethical AI solutions," commented Evgeny Lukyanchikov, Executive Director of Antiplagiat.

For reference.

System Anti-Plagiarism was created in 2005 and was developed by private individuals. The complex for checking texts for borrowings is available to both private users and has a special version for Russian universities.