Experts Claim AI Used in Drafting India's NET Exam Questions
Experts allege that AI was used to draft and translate three UGC-NET exam papers, later cancelled by the National Testing Agency (NTA). The NTA denies the accusations.
Subject experts have claimed that artificial intelligence was allegedly used in the drafting and translation of three papers for India's University Grants Commission - National Eligibility Test (UGC-NET). The National Testing Agency (NTA) cancelled these papers last month following reports of significant errors. However, the NTA has denied using generative AI to author the disputed papers.
The allegations concern the English, Commerce, and Sociology examinations held during the June 2024 UGC-NET cycle. Following earlier scrutiny, candidates flagged issues including repeated questions, misspelled names of academics, factual errors, and problematic translations. The NTA cancelled the papers after an expert committee confirmed multiple defects and has scheduled fresh examinations for September 9 and 10.
Anonymous experts speaking to The Times of India suggested that automation and AI contributed to errors that escaped adequate human scrutiny. They claimed NTA officials admitted the papers were designed in an "ad hoc manner" and that they were unable to declare the correct answer keys.
Other experts questioned the process, asking how so many questions were picked from previous years and how examiners lacked answers to their own designed questions. A JNU faculty member described a broader concern about "the degree to which automated systems were used to translate, organize and assemble high-stakes national examinations."
NTA Director General Abhishek Singh has denied the use of generative AI for authoring the papers but did not explain why the errors went unnoticed despite human review. New examinations are scheduled for September.