Generative AI a useful tool that could be making students feel lonely
A CQUniversity co-led study has found that while students are finding benefits in using artificial intelligence (AI) tools like ChatGPT, they’re not providing the sense of human ‘belonging’.
CQU Associate Professor Michael Cowling was co-chief investigator on the study which included researchers from Monash and the University of Tasmania who wanted to explore the question of whether the use of AI generative tools was exacerbating loneliness in students.
“The answer was complicated. We found that students do feel a connection to the AI in the moment when they chat with it, but that it doesn’t provide them with an innate sense of belonging,” Professor Cowling said.
“So, you feel like you’re supported at the time, but it’s not the same as being part of a real human social circle and doesn’t give the same benefits.”
The study involved surveying more than 300 students around the world on the use of AI-generative (GenAI) tools with their findings recently published in the journal, Studies in Higher Education.
“This study was really about whether or not a student talking to an AI would help them feel less lonely, very much in the vein of ‘can you make friends with a robot?’,” he said.
“This ability to treat the machine like a human is something we’ve never had before GenAI.
“Our research found that whilst there were some small indirect wellbeing benefits experienced by students using artificial intelligence for their academic studies, real belonging needed an actual human connection.”
He said the study also found that students who felt socially supported by artificial intelligence tools had poor grade performance compared to students supported by close friends and family.
One of the key recommendations of their study was the need for more discipline and training of AI to maximise the benefits to humanity and that the higher education sector needed to explore innovative ways to facilitate student belonging. Universities could be supported in this work through improved data on students’ sense of belonging, including through the Student Experience Survey.
“CQU now has a dedicated GenerativeAI Learning and Teaching page, and a university wide symposium will soon be held by Dr Meena Jha,” he added.
“But currently our focus is much more on generative AI’s impact on Learning and Teaching rather than the holistic parts relating to student wellbeing.”
Dr Joseph Crawford from University of Tasmania said: “Currently, there is a lack of training and skill in areas of prompt engineering, and probably a lack of discipline in what the AI tools should be used for, and not used for that means students who rely on these tools aren’t achieving better performance.”
Associate Professor Kelly-Ann Allen from Monash University said: “Investing in genuine human relationships is crucial for our ability to cope and thrive. Positive and satisfying relationships that take time and intentional effort to build and maintain are critical for our wellbeing and cannot yet be replicated by AI.”
The team believes that as student cohort continues to diversify and more students look to study through online or hybrid models, universities need to explore innovative ways to facilitate student belonging.
