No UGC framework, no scope of AI-free assignments; teachers rethink class assessment with viva voce

"I take a PDF and ask an artificial intelligence (AI) chatbot like ChatGPT to summarise it when I am feeling lazy," says Aadya Singh, a second-year BA LLB student at Hidayatullah National Law University (HNLU), Raipur. Poonam Verma, principal of Delhi University’s Shaheed Bhagat Singh College of Business Studies, said she has stopped expecting AI-free assignments. “AI is not the future, it's the present. Students will use it anyway whether you have a policy or not,” she said.

Some institutions are taking steps to impose some measure of discipline.

“Since the introduction of AI chatbots like ChatGPT and others, we have seen students confused, whether it's acceptable or not acceptable, when to use, how to use,” said Partha Chatterjee, dean of academics and professor of economics at Shiv Nadar University. To address this issue, the university introduced a five-step AI policy for students and teachers last April.

“The policy serves as a guide in terms of how to use and in which situations to use, what are the benefits in using and things like that. It enables people and empowers people to use it without worrying,” he said

While most universities allow 10% plagiarism in assignments and take-home projects – measured using a variety of tools – they await a regulatory framework from University Grants Commission (UGC) or universities administrations. While they await central policy to catch up, teachers at institutions like Delhi University and National Law School of India University (NLSIU) Bengaluru, have added viva voce along with written assignments and projects to gauge learning outcomes.

Also read Mofussil to Markets: SNDT Women’s University is taking fashion design boom to the Maharashtra hinterlands

Tackling AI: Viva voce, five-step policy

Chatterjee explained that the five-step AI usage policy is to balance innovation with academic integrity, allowing faculty to assign different levels of AI use where students have to disclose the level of usage of AI in course learning and individual assessments.

He explained, “At the most restrictive first level, students are largely prohibited from using AI, with limited scope for basic editing, while the second level permits its use for idea generation and structuring without contributing to core arguments. The third level allows AI for editing and refinement, often requiring students to submit original drafts, final outputs, and even prompts for transparency.

“At the fourth level, AI can be used to assist in problem-solving and task completion, provided students understand the process, while the fifth and most open level, termed responsible autonomy, enables unrestricted AI use as a companion tool, with the condition that all interactions with AI are documented.”

Sony Pellissery, professor and co-director, Centre for the Study of Social Inclusion, NLSUI stated that teachers at the premier law school have added viva voce along with assignments. “An assignment is always accompanied by a conversation with the student. The teacher sits with the student and asks a couple of questions to make sure the student understands what they have written,” he explained.

Verma’s colleagues at the Delhi college had adopted much the same strategy. “Teachers are evaluating students' assignments through a viva,” she added.

Aligarh Muslim University had included viva as part of its project requirements even before AI-use became this pervasive. “Viva wasn't introduced as an AI response at AMU but it was already the existing system, which now coincidentally serves as a check,” said Keshav Kumar, a third-year BA LLB student at AMU.

Also read Teacher Training: Deemed university on paper, NITTTRs lose ground as AICTE, MMTTCs muscle in on domain

Teachers adapting AI in assignments

Chatterjee finds making teachers adapt to AI tools especially hard.

“Change is not easy. Usage of AI in the classroom requires teachers to rethink all their teaching methods, which has been going on for years,” said Chatterjee. Some teachers favour a total ban for every assignment but that’s “a conversation that's still ongoing”. “We are holding a lot of workshops for students and faculty members because this is such a new thing,” he stated.

Kumar stated that while teachers in his university are being trained in efficient AI-use, adaptation has not been smooth or quick.

Verma stated that while some progressive universities in India have introduced disclosure norms requiring students to specify how and where AI was used, she points out the absence of a formal governance framework. “With the current issues in the classroom related to rampant usage of AI, certainly there is a lack of a governance policy and we are waiting for it to come from the university.”

Also read Promised, but missing: Five years on, National Digital University reduced to a budget item, with no funds

AI use, ethics: Future use

While universities are testing different approaches to handling AI use.

Shiv Nadar University has set up an academic integrity board to handle cases of misuse with a focus on educating people rather than penalties. “More than penalties, we have relied on trying to educate people and make them informed about the use of AI, so that they can themselves become responsible users,” Chatterjee stated. Further, the university has started an “AI companion” programme alongside regular coursework.

“Students will study both the use and the ethics of AI, including questions of intellectual property, in collaboration with practising IP lawyers. The track runs from the end of the first semester through to the final year. The intention is not to make students afraid of the tool but to make them fluent in it, and honest about it.”

Singh doesn’t favour a total ban and pointed out that detection tools are often unreliable – often, a fully original copy is declared AI-generated. Instead, she asks for training. "Students should be taught prompt writing like how you make an AI re-evaluate sources, which is missing in a lot of places."

Kumar demands that teachers should look for original ideas. “Teachers should keep a focus on originality of content. It's okay that we use ChatGPT as a guide, but what about our own original idea? They [regulators] can specify the cap of usage of 25-30% of AI tools so that academic sanctity is preserved."