This paper explores the role of Artificial Intelligence (AI) in English language assessment, focusing on the use of automarking systems to evaluate written essays and extended speech.
It examines the benefits of automarking, including faster marking speeds, on-demand testing, consistent performance, and integration of learning and assessment, whilst taking into consideration the common limitations associated with automarking.
Find out more about the principles for the responsible development and use of automarked tests, advocating for a hybrid human/machine approach to ensure fairness and quality. The paper also provides a case study on the use of automarking in Cambridge writing exams, and how the principles are implemented in operational conditions.
The research reflects Cambridge’s commitment to leveraging AI while maintaining rigorous assessment standards.
“While it undoubtedly has an increasing role to play in language assessment, it’s important to understand the limitations, constraints and vulnerabilities of automarking technology.
Its implementation should be based on research, and our study highlights the importance of blending AI with the expertise of human examiners – it really is essential to keep a ‘human in the loop’.”
Dr Jing Xu
Read the research paperRead the blog post