How university students can use AI in their studies

In the lab for his introduction to programming course, Dr Erick Purwanto is experimenting with how generative AI can support students as they learn to code.  

For example, students can prompt it to generate ideas for solving a problem they’ve encountered.  

“I don’t want the students to use AI to just generate a solution code, as they won’t learn anything from it. However, they can utilise AI to generate ideas on how to solve the problem,” Dr Purwanto said. 

Students can also use it to create multiple ‘test cases’ for checking whether their code is working. Since writing test cases is time consuming, this is a notable time-saver. 

Many coding programs will be able to alert programming students to typos or functional errors, but not logical errors. However, AI can. 

“For beginner programmers, this is where they get stuck,” he said. “And the good thing is that the AI we are using doesn’t just suggest us the corrected code, it will give us the reasoning behind it.  

“If students don’t understand the explanation, they can ask the AI to explain it in another way.” 

For the last two semesters, Angela Xia, Associate Dean for Learning and Teaching at XJTLU’s School of Languages, has offered postgraduate students a new workshop series called AI-Enhanced Literature Review.  

“I wanted to raise student awareness that you can do a review in a quite different way than the traditional method using a library database. And I wanted to show them where they could look for help,” she said.  

The series explores how AI can be used to help students find appropriate literature for their review efficiently. 

She said: “Students or researchers relying on Google, Google Scholar or library databases usually need to sift through a large amount of information, and they are not necessarily able to find what they want.” 

With AI-enhanced literature review tools, students can start by inputting one article they are interested in as a ‘seed paper’. 

“Based on the seed paper’s reference metadata, AI can create a map or a graph, and identify other articles which the algorithm considers similar,” she said. 

“This can save the researchers or students a great deal of time because they can use AI to accurately identify other papers that are relevant.” 

AI can also save time in this process by producing a clear summary of the main ideas of a paper, explaining any challenging sections and providing notes for the student to refer to. 

“If you highlight a section of the paper, you can ask the AI technology to explain it to you,” Xia said. On some platforms, the AI can explain it at varying levels of simplicity or complexity, depending on the student’s needs.