Designing an AI-resistant assignment: Can we safeguard against student use of artificial intelligence?

Artificial intelligence (AI) chatbots and large language models (LLMs) have the potential to revolutionize higher education by providing personalized learning, enhancing accessibility, and offering real-time academic support. However, ChatGPT and similar smart text-generating LLMs can be easily exploited by students to write assignments or essays, challenging traditional methods of assessment. For third-year students in the dental hygiene program at a Canadian university, a written assignment was designed by incorporating several features to prevent students from exploiting AI in their writing. A study was conducted to explore students' use of AI tools in assignment writing.

The assignment was designed as a literature review, in accordance with the milestone-based assessment approach. Qualitative and quantitative data were collected from students' written responses to reflective questions. Manifest content analysis was used to analyze the qualitative data.

Forty-seven students responded to the questions, of whom 77% (n = 36) mentioned not using AI chatbots while completing their assignments, whereas 23% reported using AI tools. The content analysis generated 3 categories, representing students' reasoning for not using AI chatbots in writing the assignment: pedagogical concerns, ethical concerns, and concerns with the technology.

As generative AI evolves, new assessment methods will be necessary to evaluate critical thinking, problem-solving, and creativity-skills that LLMs cannot replicate. The findings of this study can help educators design AI-resistant assignments while deepening their understanding of student attitudes towards and perceptions of advanced AI tools.