Letters | In the age of ChatGPT, design better tests to foil plagiarism
- Readers discuss better assessment design, and accountability for the Iraq war
The popularisation of technology like ChatGPT has understandably aroused anxiety in the education community.
Assessments are an integral part of learning and teaching. Administered properly, they provide a useful reference in education planning. Although it is possible to fend off academic dishonesty with professional judgment, it may be better to tackle the issue at source through careful assessment design.
Firstly, the proportion of closed questions should be kept to a minimum for obvious reasons. If closed questions have to be included, they should be made more complex, for example, by asking students to demonstrate inference skills. Instead of asking students to evaluate individual statements, multiple choice questions should ask them to determine whether one or more of the statements are true.
When it comes to open questions, however simple the wording is, students are expected to process information and make judgments.
The assessment should be designed to discourage regurgitation, and to promote creativity, information literacy and values education instead. Rather than turn into walking textbooks, students should be human.
Secondly, use more modalities of assessment. Other than written answers, students could submit videos. They could also be assessed based on their instant responses during online discussions on topics already covered in class.