Supporting Information S1–S10: Extracted Sections and GPT Outputs

This dataset includes the Supporting Information files S1 through S10, submitted with the manuscript titled:"Developing AI tools for IRB pre-review: A pilot study on ChatGPT’s accuracy and reproducibility."The Excel files contain original excerpts and English translations from the clinical research protocol and ICF, as well as outputs from GPTs 1–4 and standard prompts. These materials were used to evaluate the models' accuracy and reproducibility in extracting key elements such as study objectives, design, and expected benefits and risks.