ChatGPT has spread through universities faster than most lecturers could write a policy on it. Launched in November 2022, it reached 100 million users within three months, and many students now treat it as an everyday study aid. A new study from Yemen suggests that students are most ready to adopt the chatbot when it suits the work they are set and when they already think well of it.
A researcher at the Community College of Qatar and Al-Razi University in Yemen surveyed 209 students at Al-Razi University, all of whom had used ChatGPT or similar AI tools. The findings appear in The Open Psychology Journal. The online questionnaire tested five possible influences on students’ intention to use ChatGPT, drawn from two long-standing theories of why people take up new technology.
Students were recruited through social media and university student groups, and all were aged 18 or over. Just under 57% were men, about 94% were under 30, and most were undergraduates, largely studying business. Around two-thirds said they could already use ChatGPT well for learning.
Task-technology fit is the degree to which a tool’s features match the jobs a person needs to do, such as summarising reading or organising an assignment. Attitude refers to a student’s general judgement of whether using ChatGPT is a good thing. Intention describes how firmly a student plans to keep using it for coursework.
All five influences had a statistically significant positive link with intention. Students who saw their academic tasks as suited to the tool, rated its features as dependable, felt confident using it, judged it a good match for their work, and viewed it favourably all reported firmer plans to use it.
The support was not equally strong. Attitude and the nature of the tasks showed the clearest statistical links, and the fit between tool and coursework also stood out. Personal characteristics, such as confidence with the tool, only just reached significance.
Students rated statements on a five-point scale running from strongly disagree to strongly agree. They were asked, for instance, whether ChatGPT cut the time needed for difficult academic tasks, whether its answers met their expectations, and whether they would recommend it to classmates.
The author argues that universities, policymakers, and developers should build AI-supported learning around real academic tasks, keep tools easy to use, and set clear rules on plagiarism and academic integrity. Training in AI skills for both students and staff is also recommended.
The study has clear limits. It drew on one university and mostly business undergraduates, recruited through social media rather than at random. Answers were self-reported at a single point in time, so the study shows links with intention, not actual use or better grades.
Earlier research in Saudi Arabia using the same framework found that the match between tasks and technology shaped how students used ChatGPT and how well they performed. The new results point the same way, though worries about over-reliance on chatbots and weaker critical thinking remain unresolved. The author calls for studies across more universities and subjects that follow students over time.
For lecturers, the message is that students’ appetite for ChatGPT rests heavily on whether it helps with the work in front of them.
