Generative AI solutions like ChatGPT, Perplexity.ai and AutoGPT have spread like wildfire, catching the imagination of users from a wide variety of sectors. But at higher educational institutions, academic heads are worrying about how it will impact pedagogy and learning.

Although generative AI has proved to be of immense help in trawling through vast resources and quickly coming up with precise answers and summaries, it has put the academic guards in a spot. For, some students are using it to generate content for their projects. This defeats the very purpose of giving assignments, which are aimed at inculcating a habit of research and original thinking.

Disrupting learning

“We recognise the disruption a tool like ChatGPT can cause to the teaching-learning processes. We are currently exploring ways to deal with it effectively and meaningfully — both to embrace technology to help improve academic processes and take steps to curb its potential misuse,” Manoj P, Chief Operating Officer of Azim Premji University, said.

The university has begun a process from the current semester to understand how these tools are being used by students and members. The data will help it analyse and draw insights to develop strategies and plans.

The university has already started an awareness campaign among students and members of the University to prevent its unhealthy use (for instance, plagiarism) and put in place mechanisms to identify and avoid misuse.

“We are also experimenting with different types of assessments in academic courses to ensure that the influence of these tools is limited. We will find ways to meaningfully integrate tools into the teaching-learning process for the larger benefit of all stakeholders,” he says.

On the other hand, Shalabh, Professor and Dean of Academic Affairs at IIT Kanpur, asserts that Generative AI is not yet a threat to the IIT ecosystem. “The IIT curricula is a blend of theoretical as well as practical components. At IIT Kanpur, especially, our pedagogy and curricula are designed more towards developing intuitive thinking, creativity and innovative problem solving skills and are not based on text book based evaluation,” he explains.

In academic circles, there is much debate on the topic currently. G Pardha Saradhi Varma, Vice-Chancellor of KL University, says, “At a recent international meeting of Vice-Chancellors, we have discussed the challenge posed by AI and automation.” He feels that pedagogies will be impacted significantly because of Generative AI solutions. “We at KL University have decided to focus on activity-based, project-based learning and gamification,” he said. Earlier, these used to be just 10 per cent of the learning method, but now it could go as high as 60 per cent.

Understanding the threat

Rajiv Tandon, CEO of Bits Pilani WILP (Work Integrated Learning Programmes), said that ChatGPT is posing conundrums that have not been witnessed over many generations. He says there is optimism mixed with lurking fear.

It is a top priority for higher education leadership across institutions and policymakers to develop a deeper understanding of Generative AI in order to address the challenges. “Most progressive institutions are doing just that and are trying to first deepen their understanding of opportunities and implications of AI on teaching, learning, and assessments,” he said.

He warns against taking knee-jerk reactions though.

The good bits

Tandon feels there are some positive aspects too. He said there are some noteworthy efforts being witnessed especially in non-proctored assessments, such as assignments and projects.

There are ways to beat the AI, whose ability to assess and generate content is completely based on the material that is fed into it.

“In the project-based learning and assessment methods, educators are trying to move towards tasks that involve solving ‘wicked problems.’

Wicked problems are the problems that are not easily solved by just knowing the past patterns, which AI often uses to build solutions. “Many interconnected phenomena are needed to solve a wicked problem and the learner is required to demonstrate an understanding of these interconnections to design their approach to solving them,” he points out.

In addition to making more efforts for designing wicked problems for projects, some educators have also begun to make some simple yet effective changes in both project requirements and project report submission methods. Asking for a face-to-face or online live presentation on the project report can be another way to check the copycats.

Generative AI in the education system itself is a ‘wicked problem’ and institutions will need much deeper understanding and imagination to get to the right solution and use cases, Tandon sums up.