No AI guidelines for universities

Despite the widespread use of generative artificial intelligence tools among Bangladeshi university students, educational institutions have made little progress in developing policies to regulate their use, raising concerns over academic integrity and the quality of learning.

While universities worldwide have introduced rules governing AI use, Bangladesh’s University Grants Commission (UGC) has yet to issue guidelines, leaving universities to adopt differing approaches.

“AI has become a regular tool in my overall study workflow,” said Tanjil Ahmed, an MBA student at Dhaka University. He recalled recently using AI tools to complete an assignment from start to finish.

Subscriptions to major AI tools such as ChatGPT and Claude are readily available in Bangladesh. Although these tools can support learning, their growing use has raised questions about students’ independent thinking, creativity and the originality of their academic work.

“Some students are using AI to write for them. They are submitting that writing under their own names,” said A Al Mamun, professor of Mass Communication and Journalism at Rajshahi University.

The National Education Policy 2010 identifies critical thinking, creativity, original research and problem-solving as key objectives of higher education.

Concerns over education quality, however, predate AI. The World Bank’s Higher Education Quality Enhancement Project, launched in 2009, sought to improve teaching and research, while a 2014 WB overview identified inadequate funding for research, laboratory equipment, books and journals.

Prof Md Khairul Islam of DU’s Institute of Education and Research, who has researched ChatGPT use among university teachers and students, said excessive dependence on AI was affecting students’ ability to think independently and critically.

Students increasingly use AI to find research literature, summarise readings and complete assignments.

Hlathoaicha Chak, a third-year student of Persian Language and Literature at RU, said AI had fundamentally changed his reading habits. Instead of spending hours reading, he now relies on AI-generated summaries.

NM Ullash, a fourth-year journalism student at Chittagong University, said, “The constant availability of instant summaries has diminished students’ focus during lectures and eroded their inclination to engage in deep, sustained reading.”

The growing reliance on AI has also challenged traditional assessment methods, particularly take-home assignments.

Prof Imranul Hoque, chairman of Jagannath University’s Marketing Department, said teachers were encountering assignments that appeared far beyond students’ demonstrated abilities.

“But when we look at the content, we can see that, at this level, it is not possible for the student to produce something of that standard,” he said.

He warned that excessive AI dependence could weaken critical reasoning, analytical skills and independent problem-solving, and called for common national rules.

Swakkhar Shatabda, professor of Computer Science and Engineering at BRAC University, raised similar concerns.

“I have seen students submit strong or sophisticated assignments but struggle to explain or defend them,” he said.

In response, some teachers are placing greater emphasis on oral examinations, presentations, classroom discussions and project-based assessments to test students’ understanding rather than simply evaluating their submissions.

Md Abdul Kabil Khan, associate professor and head of Journalism, Media and Communication at Daffodil International University, said teachers were reconsidering marking criteria for take-home assignments, giving greater weight to students’ original ideas, creativity and knowledge.

Taspriya Labonno Progga, a student at DU’s Institute of Education and Research, said instructors were increasingly assessing students through direct classroom interaction.

“Methods such as viva examinations, presentations, class discussions, and in-class assignments are being used more,” she said. “This makes it more difficult for students to simply submit work generated by AI.”

Prof Mamun of RU suggested more oral assessments and requiring students to preserve successive drafts of assignments and projects so teachers could evaluate their reasoning and work processes.

Prof Khairul said he also asked students to present their assignments, noting that those heavily dependent on AI sometimes struggled to explain their work.

Generative AI became widely available following ChatGPT’s public release in November 2022, prompting universities worldwide to revise their academic policies.

A 2025 study, “Comparative analysis of artificial intelligence policies in universities across five countries”, examined guidelines at 343 universities in Australia, Canada, China, the UK and the US. It found that 179, or 52 percent, allowed individual instructors to determine how AI could be used in their classes.

In Bangladesh, the UGC is still preparing a national framework.

UGC Chairman Prof Mamun Ahmed said the commission planned to introduce guidelines promoting ethical and transparent AI use. Universities would then develop their own policies, with departments preparing discipline-specific rules.

“There is no reason to think that one AI policy can cover all universities, or even all departments within the same university,” he told The Daily Star.

“We hope that the guideline will be ready by January,” he said.

Meanwhile, universities are at different stages of developing their policies.

DU Pro-Vice-Chancellor (Academic) Prof Abdus Salam said a committee had been formed around two months ago to prepare AI guidelines.

Chittagong University Vice-Chancellor Prof Mohammad Al-Forkan said the university had decided to formulate an AI policy and had been organising short training programmes for students across faculties.

At Daffodil International University, revised AI guidelines covering students, teachers, researchers and administrative staff were approved in 2026, according to Md Abdul Kabil Khan.

A 2025 study co-authored by Prof Khairul, based on 765 valid responses from teachers and students in social science disciplines at public and private universities, also found differing perceptions of AI use.

Students generally considered ChatGPT helpful for learning, while teachers were more cautious about its effects. The study identified concerns over excessive dependence, unreliable information and inadequate institutional guidance on ethical use.

Prof Khairul said university policies should clearly define acceptable AI use in assignments and research for both teachers and students, while considering unequal access to devices and paid services.

He supported AI-assisted literature searches, provided students independently verified the sources.

“AI is a reality. We cannot ignore it,” he said. 

He called for AI literacy to be incorporated into university courses and for teachers to receive training in the technology’s effective and ethical use.

“There is no alternative to training teachers,” he said.