Context Engineering for Mental WellBeing Chatbots

This project explores context-engineering to bootstrap conversational agents for mental wellbeing, specifically in higher education.

Abstract

Mental health challenges in higher education, such as academic stress and social isolation, demand scalable, tailored interventions. This project explores a context engineering framework to develop mental wellbeing chatbots that leverage authentic, crowdsourced data from students in higher education.

By integrating community-driven insights (e.g. stressors, resources, demands, coping strategies) into large language models, agents may be able to deliver relevant, trustworthy support and/or companionship. The research combines a crowdsourcing interface and real-world deployment to enhance student mental wellbeing. Collaborating with psychologists, the project seeks to ensure ethical data management and measurable impact, offering a novel solution to personalise digital mental wellbeing support at scale.

Key Objectives

• Define a context engineering framework for mental wellbeing chatbots, prioritising authentic data in higher education.
• Design and validate a crowdsourcing platform to collect and process mental wellbeing context focused on higher education.
• Deploy and evaluate a chatbot across multiple institutions, assessing its impact on wellbeing and system sustainability.
• Ensure ethical data practices, scalability, and relevance through psychologist validation and student engagement.

Research Questions

• What contextual factors (e.g., stressors, demands, resources) are critical for mental wellbeing chatbots in higher education?
• How can crowdsourcing deliver authentic, high-quality data while addressing ethical and diversity challenges?
• What is the impact of a context-engineered chatbot on student mental wellbeing, and any other relevant issues such as help-seeking or stigmatisation of mental disorders?