Generative artificial intelligence in writing assessment: University professors’ perspectives and recommendations

Research has recognized the potential of Generative AI (GenAI) to support writing assessment. However, the responsible and effective application of GenAI in writing assessment remains underexplored, and consequently, university instructors are uncertain about how to integrate these tools into their assessment practices. Drawing on Mills et al.’s (2024) three-dimensional GenAI literacy framework, this qualitative study examined professors’ beliefs about incorporating GenAI into L2 writing assessment and their recommendations for responsible and effective integration. Twenty-four professors from universities in Chinese Mainland, Hong Kong, Macau, and Singapore were recruited. Data were collected through semi-structured interviews and stimulated recall tasks. The findings revealed that while professors recognized benefits of incorporating GenAI in writing assessment, they expressed concerns about certain critical issues. They emphasized the essential role of human agency and offered practical recommendations for effective, responsible, and supportive use. Based on these findings, the study proposes seven evidence-based and actionable recommendations, including enhancing instructors’ GenAI literacy, adopting multiple sources of feedback, prioritizing teacher judgment and modification for reliability, validity, and discriminative power, ensuring transparent and ethical use of GenAI, foster socio-emotional awareness, as well as establishing and implementing institutional policies and training for GenAI adoption in writing assessment. These recommendations help guide the development of institutional policies and enable university instructors of L2 writing and other disciplines to integrate GenAI into assessment practices with responsibility, ethics, and effectiveness.