Adolescence is one of the most formative stages in human development. It is a time of rapid biological, psychological, and social change, shaping not only mental health trajectories but also educational, social, and economic outcomes across the life course. Unfortunately, half of all lifetime mental health conditions begin before the age of 18, yet adolescents often remain one of the most underserved groups in health systems worldwide. In many low- and middle-income countries (LMICs), where the majority of the world’s young people live, access to mental health specialists is minimal or absent.

Against this backdrop, the University of Oxford, through its Medical Sciences Division and the MRC Doctoral Training Partnership (DTP), has launched an ambitious new research project: Conversational AI for Global Adolescent Mental Health: Evidence and Co-design with Young People. The initiative, led by Professor Mina Fazel in collaboration with Dr. Holly Bear, Dr. Mirela Zaneva, and the international NGO Girl Effect, aims to explore how conversational artificial intelligence (AI) tools can be responsibly designed, implemented, and evaluated to support the mental health of adolescents across diverse cultural and socioeconomic contexts.

This doctoral project is not just a traditional academic study—it represents a critical step toward bridging the gap between cutting-edge digital innovation and the urgent need for equitable, accessible mental health interventions.

The Global Challenge of Adolescent Mental Health

The statistics are stark: one in seven adolescents worldwide is estimated to live with a mental disorder, according to the World Health Organization. Depression, anxiety, and behavioral conditions are among the most common. If left untreated, these conditions can severely undermine educational attainment, relationships, and long-term health, contributing to cycles of disadvantage that extend well into adulthood.

In high-income countries, mental health services for young people are often stretched thin, but at least exist within health systems. In LMICs, the picture is far more concerning. There may be fewer than one child psychiatrist per million people, and stigma, lack of awareness, and financial barriers further restrict access to care.

At the same time, global adolescence is increasingly digital. Young people in both high-income and low-resource settings engage daily with smartphones, messaging platforms, and social media. These digital ecosystems represent both a risk—given exposure to cyberbullying, harmful content, and misinformation—and an opportunity: the chance to deliver support and interventions at scale in ways that are familiar, acceptable, and accessible to young people.

The Promise of Conversational AI

In this context, conversational AI—digital agents that interact through text or speech in human-like ways—offers unique potential. These systems range from simple rule-based chatbots to sophisticated AI-driven companions capable of nuanced dialogue, emotional support, and adaptive interventions.

For adolescents facing stigma, lack of local services, or a shortage of trained professionals, AI-driven conversational tools may serve as a first point of contact, offering psychoeducation, coping strategies, or guided connections to further care. Their scalability means they could reach millions of young people across geographies at relatively low cost.

Yet despite the excitement, critical questions remain. How effective are these systems? Do they work equally well across different cultural contexts? How do we ensure safety, data privacy, and ethical safeguards when working with vulnerable populations such as adolescents? And, importantly, how do we ensure that AI-driven tools complement rather than replace human care where it is available?

These are the central challenges that the Oxford-Girl Effect project aims to address.

Apply Here!

Project Overview

The doctoral project, embedded within the Oxford-MRC DTP iCASE 2026 programme, is structured around two main phases:

1. Phase 1: Evidence Synthesis and Landscape Analysis
2. Phase 2: Youth Co-design, Intervention Development, and Evaluation

The research will integrate systematic epidemiological methods, participatory approaches, and applied evaluation. By combining academic rigor with real-world implementation through the NGO Girl Effect, the project seeks to generate findings that are not only scientifically robust but also actionable for practitioners, policymakers, and technology developers.

Phase 1: Understanding the Current Landscape

The first phase of the research is dedicated to mapping what is already known about conversational AI for adolescent mental health. This includes two major strands of work:

Systematic Review and Meta-analysis

The student will conduct a systematic review of existing interventions that employ conversational agents in adolescent mental health. By synthesizing results across multiple studies, the research will aim to answer questions such as:

• Which conversational AI models have demonstrated measurable impacts on adolescent mental health outcomes?
• How do effects vary by age, gender, and cultural setting?
• What implementation strategies have proven most feasible in low-resource contexts?

This meta-analysis will provide the first comprehensive global synthesis of the effectiveness of conversational AI in this domain, filling a critical evidence gap.

Analysis of Survey Data

In addition to the review, the project will analyze large-scale survey datasets such as the OxWell Student Survey—one of the few population-level adolescent surveys worldwide to include questions on digital mental health and AI use.

By examining data on digital engagement, help-seeking behavior, and attitudes toward AI-based support, the research will shed light on patterns of acceptability and barriers to uptake. To broaden the scope beyond the UK, survey items may be adapted and fielded online in LMIC regions where Girl Effect operates, enabling cross-cultural comparison and ensuring that findings are globally relevant.

Phase 2: Youth Perspectives, Co-design, and Evaluation

The second phase of the project focuses on youth engagement and participatory research. Adolescents themselves will be central to shaping how conversational AI interventions are designed, implemented, and evaluated.

Youth Advisory Boards and Workshops

Through Girl Effect’s networks, the student will convene youth advisory boards, run participatory workshops, and conduct qualitative interviews to understand how young people perceive and interact with conversational agents. This participatory design process ensures that interventions are not imposed top-down but are developed in collaboration with those they are meant to serve.

Co-designing Interventions

Working within Girl Effect’s digital platforms—which already reach millions of young people globally—the student will co-design or adapt conversational AI tools tailored to local contexts. This may include language adaptation, cultural relevance checks, and the incorporation of youth-driven priorities.

Feasibility Studies and Small-Scale Trials

Where possible, the research may extend into feasibility studies or pilot trials, testing conversational AI interventions in real-world settings. These trials could take the form of pre-post studies, staged rollouts, or optimization experiments, providing valuable insights into both efficacy and practical implementation.

Apply Here!

Ethical Considerations and Safeguards

Working with adolescents and AI technologies raises profound ethical questions. This project will place a strong emphasis on:

• Safety and safeguarding: Ensuring AI systems do not cause harm or replace professional care where necessary.
• Data privacy: Protecting sensitive information, particularly in contexts where data misuse could endanger young people.
• Equity: Preventing the widening of digital divides and ensuring interventions are inclusive across gender, socioeconomic, and cultural lines.
• Accountability: Engaging with technology companies and policymakers to ensure clear responsibility for AI-driven interventions.

By embedding ethics at the heart of the project, the research aims to set new standards for responsible AI in adolescent mental health.

Oxford and Girl Effect: A Strategic Collaboration

The collaboration between Oxford and Girl Effect represents a unique partnership between academia and the practical world.

• Oxford’s academic team brings expertise in adolescent psychiatry, digital health, and epidemiology, ensuring methodological rigor.
• Girl Effect, as a commercial and non-profit partner, provides access to digital platforms already trusted by millions of young people, as well as expertise in participatory methods and ethical AI design.

This partnership ensures that findings will not remain in academic journals but will translate directly into practice, policy, and platform design.

Potential Impact

The project’s impact is expected to be multi-layered:

• Scientific Contribution: Filling a critical evidence gap on the efficacy and safety of conversational AI for adolescent mental health.
• Policy Influence: Providing evidence that can inform global health policy, digital governance, and AI regulation.
• Practical Innovation: Delivering co-designed, youth-centered tools that can be scaled across Girl Effect’s networks and beyond.
• Global Relevance: Ensuring applicability across both high-income and low-resource settings, advancing global health equity.

Apply Here!

Conclusion: Bridging Evidence, Technology, and Global Health

The doctoral project Conversational AI for Global Adolescent Mental Health arrives at a pivotal moment. Digital technologies are transforming societies, and adolescents are at the center of this transformation. Yet without careful design, evidence, and ethical guardrails, AI-driven interventions risk being ineffective, unsafe, or inequitable.

By combining rigorous academic methods, participatory research, and international collaboration, this project promises not only to generate knowledge but also to deliver actionable solutions. Ultimately, it aims to advance a vision where every young person, regardless of their geographical location, has access to effective, safe, and culturally relevant mental health support.

As global attention increasingly turns to the intersection of AI, health, and equity, this research represents an important step toward ensuring that technology serves as a bridge, not a barrier, to adolescent well-being.

For more information, Visit: University of Oxford, Department of Psychiatry

:opportunitiesforwomen.org/