Dr. Felicia Vulcu is an Associate Professor in the Department of Biochemistry and Biomedical Sciences, Faculty of Health Sciences, at McMaster University. Dr. Vulcu has utilized ChatGPT to assist with developing teaching activities focused on retention of students’ critical thinking and analysis skills in multiple lecture and lab based courses, beginning with scientific communications in ‘Ideas to Innovation in Biomedical Sciences’ (BIOMEDDC 3A03) and more recently moving to laboratory-based course ‘Inquiry in Biochemical Techniques’ (BIOCHEM 2L06).
Dr. Vulcu shares a concern with many faculty that students may be bypassing the thinking processes required for learning to take place by utilizing AI too early in their education. As a result, she has introduced activities into these courses that require students to analyze and correct AI generated errors.
Scientific Communications (BIOMEDDC 3A03)
For this course, Dr. Vulcu used ChatGPT to develop a fictitious science article centred on antimicrobial resistance as the basis for a critical assessment activity. The critical assessment was one of a series of assignments designed to develop students’ scientific communication skills. In this particular activity, students were asked to critically evaluate whether an article written for a general science audience communicated scientific concepts effectively while maintaining scientific accuracy, grammatical precision, clarity, and logical flow.
Creating an article that could effectively support this learning objective was an iterative and surprisingly time-intensive process. Dr. Vulcu first spent several hours researching the topic, selecting appropriate scientific literature, and identifying the key concepts she wanted represented in the article. She then developed detailed prompts specifying not only the scientific content to include, but also the types of problems she wanted deliberately embedded within the article.
Students are provided with a take home assignment where they are required to conduct an analysis of the fictious article and review for areas where the AI has created errors. The purpose of this activity is to train students in how they should be reviewing any scientific content they read throughout their degree. The activity specifically teaches critical analysis skills and engages them students in the specific learning process that can be bypassed by outsourcing work to AI. Subsequently, students were required to rewrite the article with corrections of any mistakes they identified.
Teaching Assistants mark students’ work using a rubric built around critical analysis skills and the students’ ability to correct errors. For example, they are marked on their ability to correct references, determine if the information provided is factual, and how in-depth their analysis of the article is. TAs check the rewritten version of the article to ensure students were able to accurately correct the errors.
Students then applied what they had learned through this critical analysis by rewriting the fictitious article in their own voice. They were challenged to correct the problems they had identified while improving the scientific accuracy, clarity, flow, and overall effectiveness of the article for a general science audience. A 600-word limit added another important scientific communication challenge: deciding what information was most important and communicating it clearly and concisely.
Laboratory Protocols (BIOCHEM 2L06)
In BIOCHEM 2L06, students complete a series of electronic laboratory notebook assignments designed to help them think critically about their experiments before they enter the laboratory. An important part of this process is learning how to write an effective Purpose statement. Rather than simply describing what they will do, students are encouraged to think about the bigger picture: Why are we performing this experiment? What are we trying to accomplish? What are the key experimental considerations? What results might we expect, and how will this experiment contribute to what comes next?
For the first Purpose-writing activity in the course, Dr. Vulcu uses ChatGPT to generate a deliberately flawed example for students to critique. The activity begins by introducing students to the role of a Purpose statement and the elements they should consider when writing one. Students are then provided with the AI-generated example and asked to identify what could be improved. The example contains a mixture of problems, including inaccuracies in experimental details, inappropriate scientific terminology, weaknesses in scientific writing and tone, missing experimental rationale, and incomplete consideration of safety.
The idea is simple: it can sometimes be easier to recognize problems in an existing example than to begin with a blank page. By first identifying and explaining the weaknesses in the AI-generated Purpose statement, students have an opportunity to think through what makes a strong one. They then apply that thinking by rewriting the Purpose statement themselves. This series of activities uses scaffolded activities to build from mid-level Blooms Taxonomy activities (Analyze/Evaluate) to higher level activities (Create).
These activities have provided Dr. Vulcu with a new way to incorporate critical analysis into scientific communication and laboratory-based assignments. While she has not yet formally evaluated their impact on student learning, she is currently developing surveys to gather feedback from students and teaching assistants. In particular, she is interested in exploring whether working with deliberately flawed AI-generated examples helps students become more confident in identifying scientific and communication problems and subsequently applying those skills to their own work. This feedback will be used to guide future iterations of the activities.
Dr. Vulcu plans to continue adapting these activities across her courses while also evaluating their impact on student learning. One framework she has found particularly useful is asking students to first critically analyze a deliberately flawed piece of scientific writing and then apply what they have learned by creating an improved version themselves. This scaffolded approach moves students from the Analyze/Evaluate levels of Bloom’s Taxonomy toward the higher-level task of Create.
Future iterations will introduce another interesting dimension as Dr. Vulcu’s AI use policy is more broadly implemented across her courses. Students will have clearly defined opportunities to use AI as a supportive tool in take-home assignments, raising new questions about how students may choose to use AI when critically evaluating and improving material that was itself generated by AI. These experiences, together with student and teaching assistant feedback, will help shape future versions of the activities.
—
Vulcu, F. (2026), From Critique to Creation: Using AI-generated Examples to Build Critical Thinking Skills, Teaching in the Age of AI: Examples, Retrieved from: From Critique to Creation: Using AI-generated Examples to Build Critical Thinking Skills, Licensed under Creative Commons BY-NC-SA 4.0.
Teaching in the Age of AI
Teaching in the Age of AI, Updates