Use of Generative A.I. for Biostatistics Homework

Biostatistics (BI-311) AI disclosure categories

Students report on level of AI use for BI-311 homework and independent projects. Select all that apply,

  • No AI use: I did not use any generative AI tools to complete this assignment.
  • Coding work: I used AI to help me modify R code, fix syntax errors, or troubleshoot to generate my statistical output.
  • Conceptual assistance: I used AI to help explain concepts, formulas, or the mechanics behind specific statistical tests.
  • Editing and readability: I use AI to improve grammar, clarity, or professional tone on my own written interpretations.
  • Interpretation assistance (Discouraged): I used AI to structure, phrase, or brainstorm the specific written conclusions and data interpretations for this assignment.

Note 1: Using AI for interpretation assistance is discouraged as it may lead to incorrect statistical reasoning, and pushes the envelope towards outright plagiarism. Honest disclosure is required, but rely on your own logic to protect your grade.

Acceptable and unacceptable AI use in BI-311 Biostatistics

Allowed A.I. use

Prohibited A.I. use
Concept advisor: It is OK to ask AI to explain how a statistical test works, help you interpret an R error message, or suggest which R package handles a specific biostatistical method. Uncredited coder: You are expected to execute the R code yourself within your own R environment and translate the raw output into your own words.
Coding troubleshooting: Using AI to help you modify code or fix syntax errors so you can generate your own statistical output. A.I. as Ghostwriter: The scientific narrative, logical arguments, and final conceptual takeaways of your paper must be entirely your own intellectual property. Use of A.I. to write your work is like using a paid ghostwriter to create your book, then claiming authorship.
Homework editor: Inputting your own drafted interpretation into the AI to improve grammar. Unsourced/unverified AI use: If AI is used in any capacity to assist you in analyzing data or shaping your perspectives, you are required to provide a proper citation to the source tool. Submitting AI-assisted text or interpretations without both explicit attribution and independent verification constitutes academic dishonesty.

How Biostatistics students report A.I. use

All homeworks should include a statement of A.I. use. Below is a markdown template that students can copy and paste to their R markdown script for the homework assignment.

Alternatively, update the rmarkdown template that you are using, eg, R Commander, then, at the R command line run

options(Rcmdr = list(rmd.template = "C:/path/BI311_rcmdr_template.Rmd"))

See Part 09: Making a report.

I advise placement of the transparency checklist after the student’s name and homework title.

### AI Transparency & Usage Disclosure

**Instructions:** Replace `[ ]` with `[x]` for all categories that apply to your workflow for this assignment.

**Select your level of AI use for this assignment:**
- [ ] **No AI use** 
- [ ] **Coding work**
- [ ] **Conceptual assistance.** 
- [ ] **Editing and readability.** 
- [ ] **Interpretation assistance (Discouraged).** 

**Tool & Prompt Disclosure (Required if any AI box above is checked):**
1. **Which AI tool(s) did you use?** (eg, ChatGPT, Claude, Copilot): 
* 
2. **Briefly describe what you asked the AI to do:** 
*

End markdown code.

Note 2: And yes, I asked Gemini A.I. to create a draft of this markdown template.

Justification & background

This AI use checklist was inspired by ABRCMS (Annual Biomedical Research Conference for Minority Students), which require students to disclose use of generative A.I. tools for their abstracts.

The rapid development and unstructured delivery of generative A.I., large language models, has us at a pivotal transition point in how to assist our students to achieve competency in (bio)statistics and data science in general (eg, discussion Ioannidis 2024, Telenti et al 2024). On the one hand, we know that skills develop through directed practice with opportunities to generalize on a variety of data sets gathered within many experimental and observational backgrounds (Ericcson et al 1993). Thus, it is best if students forgo any A.I. assistance — work with the materials recommended by the instructor, read original research papers, learn by modeling approaches used by leaders in the field, with instructors adopting project-based learning (cf discussion about “never-skilling,” Ke et al 2026). I don’t think that will change in the near future as best practice.

Ideal, perhaps, but a hopeless effort which risks setting students up for attempting to deceive instructors. It also ignores how professionals will likely incorporate AI in their own research and data science projects. Moreover, I find myself working with AI tools more and more to help me create and modify coding work, from R snipppets to my project developing LibreOffice extensions to run R in Calc as an alternative to R Commander (currently in development stage pending release in Fall 2027). Thus, I am exploring how best to incorporate use of AI in teaching science courses, and particularly, for courses that engage with data analysis. I think transparency is the best we can do.

References

Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406.

Ioannidis, J. P. A. (2024). 2024 Association of American Physicians Presidential Address. Transparency, bias, and reproducibility across science: A meta-research view. The Journal of Clinical Investigation, 134(22).

Ke, Y., Jin, L., Ong, J. C. L., Thirunavukarasu, A. J., Car, J., Cheung, C. Y., Tham, Y. C., Ting, D. S. W., Ong, M. E. H., Compton, S., Narayan, A., Keane, P. A., Wong, T. Y., Bates, D. W., Tan, P., & Liu, N. (2026). AI-induced never-skilling in medical education. Nature Medicine, 32(6), 1997–2006.

Telenti, A., Auli, M., Hie, B. L., Maher, C., Saria, S., & Ioannidis, J. P. A. (2024). Large language models for science and medicine. European Journal of Clinical Investigation, 54(6), e14183.