Episodic Future Thinking Chatbot

Files

AIChatbotWebsitePresentation.pdf (1.44 MB)
Downloads: 8

AIChatbotWebsitePresentation.pptx (1.82 MB)
Downloads: 6

AIChatbotWebsiteReport.pdf (997.55 KB)
Downloads: 16

AIChatbotWebsiteReport.docx (1.64 MB)
Downloads: 5

TR Number

Date

2024-05-09

Authors

Journal Title

Journal ISSN

Volume Title

Publisher

Virginia Tech

Abstract

In healthcare, addressing lifestyle diseases such as obesity and type 2 diabetes through innovative methods is crucial. This project introduces a user interface (UI) for an artificial intelligence (AI) chatbot designed to enhance episodic future thinking (EFT). EFT encourages patients to visualize future scenarios, aiding in decision-making processes that favor long-term health benefits over immediate pleasures. Utilizing AI, this chatbot aims to deliver accessible, high-quality care, helping users focus on significant health goals.

Initially, the project intended to adapt an existing codebase for UI development. However, it became clear that creating a new UI from scratch would better meet our specific needs. This new interface includes four main pages: Login, Chatbot Interaction, Usability Assessment, and an About page. Each page is carefully crafted using React JavaScript to ensure dynamic user interactions, complemented by Cascading Style Sheets for aesthetic design. Python is used to facilitate connectivity with the client’s chatbot backend and to manage a database that stores user information.

Users begin their journey on the website by logging in with credentials obtained after completing a demographic survey via Qualtrics, where they also consent to participate in the study. Subsequently, users interact with a GPT-4 powered chatbot, which guides them through personalized future-thinking scenarios based on their inputs. After the interaction, users assess the vividness of their imagined scenarios and the quality of their chat experience, rating their responses. This data, along with chat logs, is stored for analysis and further enhancements. Finally, participants provide detailed feedback through a structured numerical assessment, contributing to continuous improvement of the chatbot’s effectiveness.

Through a meticulous design and iterative testing process, this project not only addresses critical health issues but also sets a new standard for integrating AI with user-centric healthcare solutions.

Description

AIChatbotWebsiteReport.docx This document provides a comprehensive overview of the development process, use cases, and the functionalities implemented in the Episodic Future Thinking (EFT) chatbot user interface. It details the design decisions, user interaction models, and the technological framework utilized throughout the project. AIChatbotWebsitePresentation.pdf This presentation outlines the motivations behind the EFT chatbot project, details the specific requirements, and summarizes the completed work on the user interface. Additionally, it provides insights into the database structure and the operational flow of the site post-credential acquisition from Qualtrics. AIChatbotWebsitePresentation.pptx Similar to the PDF version, this PowerPoint presentation emphasizes the motivations, requirements, and the developmental accomplishments of the EFT chatbot's user interface. It also illustrates the database structure and navigational flow of the site after users obtain their credentials via Qualtrics. AIChatbotWebsiteReport.pdf This report documents the developmental journey, relevant use cases, and the comprehensive work accomplished on the EFT chatbot user interface. It elaborates on the project's scope, user engagement strategies, and technical infrastructure.

Keywords

Episodic Future Thinking, EFT, GPT-4, Website

Citation

Persistent link

https://hdl.handle.net/10919/118740

Collections

CS4624: Multimedia, Hypertext, and Information Access