The UGC-NET examination has been under renewed scrutiny in the wake of allegations of artificial intelligence in the preparation and translation of question papers, inadequate human review and a long delay in cancelling three papers that are subject to serious errors. The debate has largely been about the English, commerce and sociology exams for which the National Testing Agency (NTA) has ordered new tests.
According to reports, subject experts and NTA insiders say artificial intelligence was used extensively in the preparation of some examination papers. But NTA Director General Abhishek Singh has denied that generative AI was used to write the disputed question papers. The contradictory accounts have also fuelled a debate on the role technology is playing in high-stakes examinations and what level of security is required if automated tools are used.
Why AI use in UGC-NET paper preparation?
The main issue is not simply whether AI was used to prepare the papers, but whether enough academic and human oversight was provided after any automated processes were completed. Analysts who have been quoted in these reports claimed that AI was involved in preparing and translating examination material and called the alleged implementation “reckless use of AI”.
The former subject experts who worked on the examination system said similar technology had been used to translate NEET-UG papers. Such allegations raised questions about whether automated systems were being used for tasks that would require subject-specific accuracy in exams where even a small error may affect thousands of candidates.
NTA has maintained that generative AI was not used to write the disputed question papers. The agency has also asserted that human review was done after the papers were created. But questions remain about the extent and quality of that review, especially given the number of mistakes reported by candidates and experts.
Errors reported in Sociology and English Papers
The sociology paper became a huge source of complaints with candidates citing misspelled names, distorted terminology, awkward translations and questions which were apparently outside the prescribed syllabus. Names of major sociologists and academics were reported to be incorrect, raising questions about the paper-setting and translation processes.
The English paper also faced allegations of extensive repetition. It was reported that nearly 67 of the 150 questions had appeared in an earlier UGC-NET examination. The repetition has raised the question of how questions are chosen and reviewed before a national-level examination is taken.
Commerce candidates also reported repeated questions that included taxation, break-even analysis, capital structure and services marketing. While some overlap may be present in competitive examinations, the extent of repetition has raised questions about the originality and quality-control mechanisms employed by the examination authority.
Lack of Human Scrutiny under the Spotlight.
The controversy has also brought into focus the role of subject experts in the paper-setting process. A senior Education Ministry source reported that in conversation with senior NTA officials it was not possible to finalise answer keys due to the papers being designed in an “ad hoc” manner.
There were also questions about why the papers were not subjected to stronger expert review before being administered. If a panel of subject specialists had been involved, experts have argued, the obvious errors and repeated questions could have been thoroughly reviewed in multiple stages.
But the issue is particularly important because UGC-NET is not an ordinary examination. The test determines eligibility for teaching and research opportunities in Indian universities. Errors in the question paper can have consequences for candidates' academic and professional careers.
Six-week delay in cancelling Flawed Papers.
The timing of NTA's response has added another layer to the controversy. Education Ministry officials and NTA sources had been thinking of cancelling the three affected examinations and performing fresh tests by the end of July or the beginning of August, the source said.
However, the cancellations were announced nearly six weeks after the discrepancies emerged. At that time, provisional answer keys had been released for 84 of the 87 subjects.
The delay has raised questions about whether the concerns could have been addressed earlier and whether candidates were given timely information about the status of the examinations.
Bigger Questions About AI In High-Stakes Exams
The UGC-NET controversy has exposed a larger issue for examination bodies as AI is increasingly prevalent in education and administrative functions. AI tools can help in translation, formatting, question organisation and other routine tasks, but use in high-stakes examinations needs to be carefully regulated.
As subject experts have said, automated systems can’t replace experienced academic reviewers, especially when questions are about certain terminology, names, theories and concepts. Any AI-driven workflow needs to go through a number of layers of verification to ensure that the final examination paper is accurate, syllabus-compliant and free of ambiguities.
The controversy has raised several questions: who approved AI-based systems? Which tasks were automated? Were qualified subject experts involved at every stage? How thoroughly were the final papers reviewed? And why did it take several weeks to cancel examinations which were reported to have significant problems?
For UGC-NET candidates, the immediate priority is a fair and reliable examination process. The episode has only strengthened the need for transparency, accountability and human oversight in national-level competitive examinations.
