KOLKATA: Calcutta University (CU) is set to introduce new regulations for PhD research that will restrict the use of artificial intelligence (AI) in the preparation of research papers and dissertations. The proposed standards would allow no more than 10 percent of doctoral theses to be written using AI.
“Details will be made public after approval at the upcoming syndicate meeting,” a senior CU official said.
University officials said CU already uses University Grants Commission (UGC)-approved plagiarism detection software to check copying from existing research papers and books. The university is currently planning to deploy specialized software to detect the extent of AI use in academics. If more than 10% of your research paper or paper is found to be created using AI, your submission will be rejected.
“We already have a system for reviewing research papers. We are currently introducing controls on the use of artificial intelligence and are upgrading the software for this purpose. The syndicate will take the final decision,” said another CU official.
CU will be the first university in the state to implement such specific measures regarding the use of AI. Several universities across the country have already introduced similar systems, subject to conditions.
Sanatan Chatterjee, general secretary of the Calcutta University Teachers’ Association, said the trend of researchers using AI tools available online to prepare papers is of serious concern. “This behavior is extremely harmful to higher education and needs to be regulated. That’s why universities have taken up this issue. The upper limit for such plagiarism should be less than 10 percent,” he said.
Indian Institute of Technology Bombay and Indian Institute of Technology Kanpur allow the use of AI in coding and technical work only if researchers clearly justify its use. The Indian Institute of Science treats the use of AI to generate images as scientific misconduct.
Jawaharlal Nehru University and Delhi University use specialized software to identify AI-generated text and reject papers that exceed their AI usage
10 percent.
