In-app interventions and LLM-generaterd misinformation

Overview

This project aims to study how effective various in-app interventions are at improving humans' ability to detect health misinformation generated by large language models (LLMs).

Reproduce the results

To reproduce the results, please install R and python.

1. Run results_final.r in R to generate all the necessary odds ratios and confidence intervals.
2. Install the relevant python packages in requirements.txt using pip install -r requirements.txt.
3. Run plots_final.ipynb in Jupyter Notebook to generate the plots.

Check the figures folder for the generated plots.