AI can lighten teaching load if lecturers determine its tasks

Artificial intelligence (AI) could help university lecturers reduce their increasing teaching workload as massification continues to intensify, especially at public universities. However, AI adoption is hindered by factors including a lack of policy guidelines, funds and skills, a Ugandan study shows.

Using AI to manage workload can only be effective when lecturers dictate which tasks AI can manage to erase the fear that it could replace them, the study, conducted at five public universities, found.

In an article published in the East African Journal of Education Studies on 24 July 2024, Dr Edith Namutebi of the Mountains of the Moon University, Uganda, writes that the student population in Uganda’s 13 public and 39 private universities increased almost four times to 198,000 in 2022, up from an enrolment of close to 50,000 in 2000. The higher rates are found in public universities.

Namutebi says the higher enrolment has not only “overstretched” the physical instructional infrastructure of public universities such as Makerere and Kyambogo, but also the teaching capacity of academic staff members.

Overstretched lecturers do not grow

Whereas most of Uganda’s public universities meet the recommended National Council for Higher Education’s (NCHE) faculty-student ratio of 1:25, the researcher argues that “this ratio applies only when the total of lecturers is compared to the total enrolment”.

For instance, Namutebi cites Makerere’s average total enrolment of 30,574 students and 1,274 faculty members, which gives a faculty-to-student ratio of 1:24. However, it surpasses the NCHE’s 1:25 ratio when computed per course unit, especially at the undergraduate level in arts, social sciences, natural sciences, and education, where some classes have more than 300 students.

Namutebi says the country is recording excessive undergraduate enrolments “yet the corresponding academic staff size is very small, leading to far more than the recommended lecturer-student ratio, and coming in with excessive teaching workloads that overstretch lecturers”. This, she says, has led to limited research by lecturers, stagnant career growth, and almost no community engagement.

The researcher interviewed 325 lecturers, including faculty deans and heads of departments randomly selected from five public universities, to analyse their awareness of the teaching tasks AI can execute to reduce the workload without replacing them, the level of acceptance of AI to perform these tasks, and hindrances to its adoption.

AI should not replace human touch

According to the study, AI can be leveraged in handling all teaching tasks executed using hard skills to minimise lecturer facilitation efforts and related financial resources and maximise the impact. Most of the respondents (74%) were highly aware of the teaching tasks AI can perform. They agree that AI should be adopted to perform teaching tasks that “do not involve a human touch as an online search for research and lecture content, lecture dictation, student assessment and evaluation, and grading of marks”.

The respondents, however, did not accept AI to execute teaching tasks that involve the human touch, especially lesson planning, facilitating tutorials and discussions, assessing students’ interpersonal weaknesses that affect learning, and feedback provision.

“These findings allude to a need to adopt AI to execute only the teaching tasks it is accepted to perform and leave to the lecturers all the tasks they do not accept AI to perform. Adopting AI this way is bound to relieve the teaching workload allocated to lecturers as massification intensifies,” Namutebi writes.

Other factors hinder adoption

She told University World News that lecturers can be left to facilitate teaching tasks that need human and soft skills like lecture planning, facilitating tutorials and discussions, assessing students’ interpersonal weaknesses that affect learning, understanding students’ emotional state, being responsive to changes in the classroom attitude, controlling social interruptions in lecture rooms, being responsive to spontaneous questions students can pose during an ongoing lecture, subjective marking of essays and awarding of marks in course works, tests and examinations.

This is because “AI is designed to handle routine teaching tasks that require hard skills like automatic searching for teaching content online, logical/systematic and uninterrupted teaching and automatic marking when answers are in-built and definitive,” Namutebi said.

However, the study found that AI adoption in public universities in Uganda is hindered by a lack of strategic, ethical, and policy guidelines, a lack of finances, and lecturers who are ill-equipped to work with AI.

In the article, Namutebi urges the management of Uganda’s public universities “to adopt AI by lobbying the government for more funding, mobilising the necessary funds internally, training faculty members in using AI, and encouraging all of them to accept it by explaining the role it is capable of playing in reducing workloads and erasing their fear that AI could replace them”.