NEW DELHI: Indraprastha Institute of Information Technology (IIIT-Delhi) has allowed the use of Artificial Intelligence tools such as ChatGPT in examinations and assignments as part of a major overhaul of its evaluation system. Under the new model—currently being piloted in select departments—students must submit the prompts they used along with their answers. The institute already conducts open-book exams in some courses. Officials said the move aims to bring teaching and assessment practices in line with technological shifts and evolving learning patterns. Discussions are underway to develop a uniform framework for undergraduate, postgraduate and doctoral programmes.
As proposed, some tests may be held as take-home assessments, while others could permit students to access AI tools on their laptops during supervised on-campus exams. Faculty members piloting the model said evaluation will focus on the quality of prompts, reasoning and analysis rather than memory-based responses. IIIT-Delhi director Ranjan Bose said the decision reflects the institute’s effort to integrate AI into education rather than restrict it. “Students are already using AI tools in some form. The question is not whether to allow them, but how to help students use them responsibly and creatively,” he said, adding, “Originality will lie in how they frame and refine prompts to obtain meaningful outputs.” The institute has also introduced mandatory disclosure of AI use in thesis and research work. Students must sign a declaration taking responsibility for academic honesty and transparency. Additionally, all students will attend a compulsory session at the start of each semester on the ethical and academic use of AI. Bose said faculty members will also undergo training to design questions that test higher-order thinking. “We already conduct open-book exams where students are assessed on their understanding rather than their ability to recall facts. This model builds on that principle,” he said. “It will require teachers to frame questions that test comprehension and judgement, not just the ability to generate an answer,” he added. Faculty involved in early trials said the institute’s approach is rooted in transparency. “Instead of banning AI, we are asking students to show how they used it,” said a senior professor. “Submitting prompts makes their process visible and helps us evaluate intent and reasoning, not just the final output,” the professor added. Students said the proposal acknowledges current learning practices. “We already use AI tools to check concepts or summarise readings. Bringing it into the exam system openly removes the stigma and helps us learn responsible use,” said Tanisha Mittal, a first year B.Tech student. Another first year student of B.Tech, Aryan Kumar, said the move could help shift focus away from rote learning. “If we’re evaluated on how we ask questions to AI and interpret its responses, it tests understanding far better than memorisation,” he said. IIIT-Delhi, which has around 3,000 students across 32 programmes, is expected to roll out the new model in the coming months. Officials said the framework will be refined based on feedback from ongoing trials before being implemented across all departments. The move comes as universities worldwide grapple with how to address AI-generated content. While many institutions impose restrictions, others are adopting structured inclusion with mandatory disclosure. Within Delhi, IIIT’s framework is among the first to formally integrate AI into the learning and assessment process.
