GenAI generated exemplars to improve student scientific writing
This project explored whether using generative AI as a writing tool could support students in Biology and Biomedical Sciences courses. Through a combination of three surveys, and interviews, we examined if AI helps reduce writing-related anxiety. Furthermore, we investigated which parts of the writing process students seek AI’s support for. Results show that AI can reduce anxiety, especially among younger students, but GenAI adoption has to be critically implemented.
Background information
We observe that students find it difficult to structure their writing logically and/or correctly to make a fluent story. The writing assignment also causes anxiety and sometimes a writing block. Exemplars (example texts generated by earlier students or based on textbooks or scientific publications) help reduce this stress (Yucela et al., 2014), but given the variety of topics in the assignment, generic examples are often not relevant. The use of generative Artificial Intelligence platforms (GenAI) as a writing tool will be more common and accepted in the near future. This project explored whether its use could lower student anxiety, and identified which writing tasks students found most suited for GenAI support. These insights guide the responsible integration of AI in writing education.
The learning mechanisms that genAI use could tap into are:
Regulative: The use of GenAI to generate exemplars could give students a start with structuring a text, help them write grammatically correct text or use GenAI as a feedback tool. In comparing their own work with the GenAI generated text, students can make improvements and self-regulate their learning, and identify which parts they find challenging or where they want to seek support.
Cognitive: Students can use AI to explain complex primary literature in molecular and cell biology, and when unfamiliar techniques are involved. Also, GenAI may support students by providing examples of scientific writing, which can help them structure their thoughts and improve clarity.
Motivational: The option to use AI after the initial instruction may reduce anxiety, potentially increasing students’ confidence and engagement.
In this contribution we focus on how genAI may reduce anxiety, and for which parts of the writing process students would use genAI.
The study
Biology and Biomedical students in the Bachelor’s course Developmental Biology had to write a mini review. They first received a lecture instruction on GenAI’s possibilities, limitations, and its responsible use (see link to material below). Then they were allowed to use GenAI to help them. They were free to choose how, to help them to produce text, or as a feedback tool. They could not directly copy+paste genAI suggestions: they had to discuss GenAI’s suggestions amongst each other in their group, critically examine them, and rewrite the text before turning it in as a final version.
We conducted the study in 2 course periods during the 2023-2024 academic year. In period 1, 145 students participated in the course as part of their bachelor’s in Biology, in period 4, 90 students participated in the course as part of their bachelor’s Biomedical Sciences. We conducted 3 surveys (see table 1) per course.
First, students received an instruction of what is expected for writing a mini review. After this instruction, students were asked to complete the first survey: a Spielberger State-Trait Anxiety Inventory (STAI) to assess their concerns about the assignment (Marteau and Bekker, 1992) and questions to identify which part of the writing process worried them. After the genAI instruction lecture, a second survey was completed with the same STAI items and questions on intended AI use. After submission of the final assignment, a third STAI survey and follow-up interviews were conducted. Students who used AI reported which parts it helped with, while non-users shared their reasons.
Table 1: The surveys
The results
Analysis of student responses showed that permission to use AI tools reduced writing-related anxiety, especially among Biology students at the start of their second year. Among biology and BMS students who completed all three surveys (number of respondents = 15), there was a significant drop on scores on STAI statements like “I am tense” and “I am worried.”
A similar trend was seen in biology students who completed the first two questionnaires (n = 43; irrespective of if they completed survey 3 or not). However, this effect was not observed in Biomedical Sciences students (n = 50), likely due to their greater writing experience at the end of the second study year.
Among all respondents (n = 93, period 1 and 4), students primarily sought AI support for:
- Writing in proper academic English.
- Understanding, reading, and summarizing primary scientific literature.
- Composing a strong discussion section.
Open-ended responses on concerns on AI use are:
- Acknowledging that they considered using AI due to a high workload, but this would not teach them writing.
- Fear of improper AI use by teammates.
AI tools can contribute to reduced student anxiety in writing assignments, particularly in young students. While students generally appreciated the instruction on AI use, students’ concerns about AI’s effectiveness, ethical considerations, and potential misuse persist. The results help us to shape a teaching strategy, ensuring that students learn to use AI responsibly, while maintaining academic integrity by instructing to check the AI generated content critically and not allowing to copy+paste AI generated text
Next steps
The use of AI in learning and helping in assignments will become common. By mentioning AI in our lecture instruction clearly as a support tool not as a final product, and by active discussion of genAI output in our tutorials, , we hope to stimulate deeper student reflection and engagement in the writing process. We plan to place more emphasis on teaching effective prompt design so students can use GenAI for higher-order writing skills. AI could be used to help students better understand complex scientific literature. We also want to explore how AI can support the revision process. Continued teacher guidance remains essential, particularly in ethical use, critical thinking, and academic integrity.
Lessons learned Tips
Attachments
Instruction AI use in writing assignment Developmental Biology
Survey 1_SoTL_GenAI_scientific writing_bachelor Biologie_voor AI instructie
Survey 2_SoTL_GenAI_scientific writing_bachelor Biologie_na AI instructie
Survey 3_SoTL_GenAI_scientific writing_bachelor Biologie_na schrijven
Central AI policy All AI-related activities on this page must be implemented in line with Utrecht University’s central AI policy and ethical code.
Responsibility for appropriate tool choice, data protection, transparency, and assessment use remains with the instructor.