Perceptions of Generative AI in Formative and Summative

Abstract

The emergence and rapid evolution of Generative Artificial Intelligence (AI) technologies, extensive large language models (LLMs) such as ChatGPT, Claude, and Gemini, are reshaping the landscape of educational assessment. These tools are increasingly utilized to deliver scalable, personalized feedback and to streamline both formative and summative evaluation processes. This study investigates the dual promise and potential pitfalls of incorporating generative AI into assessment practices within Nigerian higher education institutions. Drawing on a quantitative research approach, data were gathered through a structured Google Form completed by 140 respondents comprising both educators and students. The findings reveal a widespread adoption of generative AI tools, with 86.1% of participants indicating prior use in teaching or learning contexts. While respondents acknowledged the value of AI in enhancing the efficiency and responsiveness of assessment and feedback mechanisms, they also expressed significant concerns regarding fairness, algorithmic bias, ethical implementation, and the risk of diminishing the role of human judgment. These insights underscore the urgent need for transparent, ethically grounded, and inclusive policy frameworks to guide the responsible integration of AI technologies into educational assessment practices.