A recent study conducted by researchers at Cambridge University has shed light on the key features of essays generated by artificial intelligence (AI) technology. The study compared essays written by three university students assisted by ChatGPT to 164 essays from GCSE students.
According to the study, AI-written essays exhibit distinct characteristics such as repetition of words and phrases, the common use of introductory phrases like ‘however’, ‘overall’, or ‘moreover,’ as well as the inclusion of numbered lists and complex vocabulary.
Interestingly, the research found that essays generated with ChatGPT struggled in analysis and comparison skills compared to human-written essays. However, the AI-assisted essays performed well in terms of providing information and reflecting on the given topics.
This study has prompted university departments to review their policies on AI technology usage, especially in light of the upcoming exam season. Currently, Cambridge University prohibits the use of AI technology in assessed work due to academic misconduct concerns. However, the guidelines for non-assessed work vary across different departments.
In response to the study’s findings, students are advised to exercise caution when using AI tools for writing tasks. While grammar and spelling correction tools are allowed, those that actually write or edit content on behalf of students are considered inappropriate.
The debate on AI usage in academic settings continues, with different departments providing varying levels of guidance to students. Engineering students, for example, are allowed to use ChatGPT for structuring coursework as long as they disclose its usage and provide information on the prompts used.
As universities navigate the evolving landscape of AI technology in education, the balance between leveraging its benefits and upholding academic integrity remains a central concern. The guidance provided to students reflects ongoing efforts to define the boundaries of acceptable AI usage while preserving the essence of learning and academic rigor.
