Tools like ChatGPT have been attributed for their ability to explain concepts, write code, summarise information and save time in various ways.
Whilst academic institutions in the UK have largely acknowledged the new presence AI has in the student workflow, the HEPI study shows that 18% of students admitted to using AI to directly produce their assessed work.
Most higher education institutions in the UK have implemented academic rules to ban AI in assessed work. In order to enforce this, software designed to detect AI text has become widely used by universities, especially in the humanities.
Despite this, some academics have expressed concerns that AI could threaten academic rigour and quality in the long run. Some of these concerns are based on the reliability of AI detection tools in assessed work being challenged by ever more sophisticated large language models able to fool these tools and remain undetected.
Some students are also developing strategies in a bid to thwart the detection tools; interleaving human written sentences within AI paragraphs as well as adding uncommon and unusual words into AI generated text.
The HEPI study also shows that students from higher socio-economic groups are more likely to use AI whilst at university than students from lower socio-economic groups.
Currently, only 36% of students say they have received support from their institutions to develop AI skills. This raises questions about if academic institutions could do more to ensure all students are able to benefit from AI equally.