An Investigation Into ChatGPT Generated Assessments: Can We Tell the Difference?


The impact of ChatGPT has been revolutionary in many capacities however institutions are beginning to see the gradual increase in students passing off AI generated work as their own. This has negative impacts for student learning and academic integrity. One way to help combat this is to understand if we can tell the difference between AI generated assignments and original pieces of work. This will help those involved in assessing to distinguish between AI generated work compared to original work. In the initial phase of this study, we use ChatGPT to generate assessments for 3 modules in the department of Biochemical Engineering, UCL. These assignments capture the interdisciplinary nature of Biochemical Engineering as well as the diversity in assignment complexity and include mathematics, business and bioprocess validation and quality control. We then convene academic leads and marking staff to assess scripts, compare them to previous cohorts through use of peer-observation to find out what indicators there are of generated work. Results so far have shown that out of the 3modules, 2 modules receive a pass mark with minimal prompts and only 1 module lead was able provide indicators to identify generated work. Results also show that ChatGPT was unable to provide solutions for complex mathematical problems, bioprocess piping and instrumentation technical drawings and critical analysis required for M-level bioprocess quality control. Subsequent phases of the study will expand the number of modules tested on ChatGPT, embed its use into the engineering curriculum and upskill academics on the use of AI tools.

Author Information
Mauryn C. Nweke, University College London, United Kingdom
Matthew Banner, University College London, United Kingdom
Manal Chaib, University College London, United Kingdom

Paper Information
Conference: BCE2023
Stream: Design

This paper is part of the BCE2023 Conference Proceedings (View)
Full Paper
View / Download the full paper in a new tab/window

To cite this article:
Nweke M., Banner M., & Chaib M. (2023) An Investigation Into ChatGPT Generated Assessments: Can We Tell the Difference? ISSN: 2435-9467 – The Barcelona Conference on Education 2023: Official Conference Proceedings
To link to this article:

Virtual Presentation

Comments & Feedback

Place a comment using your LinkedIn profile


Share on activity feed

Powered by WP LinkPress

Share this Research

Posted by James Alexander Gordon