Assessment design in an AI world

The rapid evolution of generative AI has profound implications for what we teach, how we teach it, what and how we assess, and how students engage with their curriculum and assessments. We know from both our own research, and reports from across the sector, that students’ practices around use of AI in the process of planning and producing assessed work vary significantly and are influenced by a complex array of factors. These include discipline, type and nature of the task, sense of engagement and belonging on the course, and time constraints. Recent research from WONKHE has highlighted that how and whether students use AI is also crucially influenced by whether they believe they will need to demonstrate their own (unassisted) knowledge or skills in that area live, either during their studies or in graduate employment. 

The University’s guidance on the use of AI in assessment establishes 4 categories of permitted AI use, from no use to AI-integrated assessment, to enable staff to design assessments which permit appropriate use of AI for their task and context, and to communicate those expectations clearly to students. However, we know that many staff are concerned about the impact of AI on academic integrity. There are considerable challenges in restricting or disincentivising student use of AI to only some permitted purposes, or in establishing how AI has been used. In this context, it is desirable to redesign assessments to be more AI-robust, whether that is through more fully AI-integrated tasks, those where AI doesn’t confer significant advantage, or assessments incorporating some kind of in-person test of students’ skills and knowledge. 

In that assessment redesign process, it is important that the new assessment still conforms to the three priorities of the assessment and feedback strategy: integrated, designed for all, and authentic. 

Current activities

BILT’s involvement with this theme includes:

  • Funding of 4 Education Development Projects
  • Engagement of a BILT Student Fellow (Samita Khondokar) in the theme area 
  • Show, Tell and Talk workshops

Resources and guidance