Perspectives on text-generative AI in Economic assessments

Our aim is to feed into the wider discussions on the use of artificial intelligence (AI) in higher education and specifically in assessment. The debate on the use of generative AI is raging and new platforms are appearing quickly, such as GPT4 as a successor of ChatGPT, BingAI, etc. The tension between traditional and innovative forms of assessments, as well as considerations of ethics and integrity, are prominent in day-to-day life of the academics, however, a clear consensus has not yet emerged.

We survey module leaders teaching various economics disciplines on their philosophical-pedagogical approach concerning the use of essays as a summative assessment. We present colleagues with a case study from a Microeconomics module, where we ask them to find the needle in the haystack, i.e. identify the answers generated by the AI from the mix of actual students’ scripts. If the distinction was clear, then traditional assessments may still beviewed as an effective tool for evaluating students’ performance. Alternatively, we investigate whether academics consider necessary to propose an assessment re-design in the short run, whether any adjustments to the marking criteria will be made, or whether more structural suggestions are emerging, such as educating students on how to use generative AI as a new tool in an ethical manner.