Media release: Researchers say universities need to redesign curriculums after ChatGPT aces Test of Understanding in College Economics

Researchers from the University of Melbourne, University of Texas and Virginia Tech say educational institutions need to redesign curriculums after ChatGPT performed better than 90% of students in one of the most popular standardised tests for economics.

The researchers asked ChatGPT to complete the Test of Understanding in College Economics (TUCE) and found it ranked in the 91st percentile for Microeconomics and the 99th percentile for Macroeconomics.

Researcher Dr Nikhil Damodaran from the Department of Economics, University of Melbourne, said institutions need to proactively take steps to prepare students for working alongside AI in the field of economics.

“It’s clear traditional curriculums are outdated and need to be redesigned to address the needs of the next cohorts of students. As AI technology becomes more available and smarter, educators need to ensure a balance of hands-on learning and helping students understand a technology that will be present in their future jobs,“ said Dr Damodaran.

“It is important for instructors to facilitate student training so that their skills supplement the use of AI rather than compete with it,” added Dr Damodaran.

The researchers say institutions will need to redesign their curriculum in at least one of three ways:

  1. Reintroduce proctored, in-person assessments: Reintroducing in-person assessments under controlled conditions ensures students’ knowledge is rigorously tested without assistance from AI.
  1. Augment learning with chatbots: Including AI like ChatGPT in teaching can enhance student engagement and provide personalised assistance.
  1. Embrace experiential learning projects: By including experiences such as classroom demonstrations, experiments and case studies, educators can foster critical thinking, problem-solving, and creativity, which are areas where AI is limited.

See the full research results in The American Economist.

The authors were Professor Wayne Geerling and Professor G. Dirk Mateer from the University of Texas at Austin, Professor Jadrian Wooten from Virginia Tech, and Dr Nikhil Damodran from the University of Melbourne.

Media enquiries

Contact: FBE-media@unimelb.edu.au | +61 3 8344 5793

AI robot taking exams
Source: Dall E 3