Nowshin-Ara-Nini/Ai-Dependence ? reverse-engineered prompt

Build me a small Streamlit web app for a university survey project about AI dependence and cognitive offloading among students in Bangladesh.

I want the app to let someone answer a set of demographics and questionnaire questions about AI use, trust, offloading, and academic decision making, then show an estimated AI Dependency Index with a short plain language explanation that this is only an exploratory estimate, not a diagnosis or a measure of intelligence. Please make the form feel clean and simple, and keep the five dependency questions out of the inputs so there is no obvious target leakage.

It should use the saved model pipeline from the project, and also include a few summary visuals or results from the analysis if that fits naturally. If needed, look up the current Streamlit and scikit learn docs online so the app is set up the right way. I’d also like the code to be organized so the notebook analysis, the model training, and the demo app all make sense together, with a polished landing page and clear disclaimer text.