Generative AI for anxiety, depression and psychosis

Objectives:

• Fund fundamental research using generative AI to improve measurement or treatment of anxiety, depression, and psychosis.
• Optimize generative AI models specifically for mental health applications.
• Explore safe and efficient collaboration between generative AI, mental health professionals, and people affected by these conditions.
• Lay groundwork for a new generation of mental-health-first generative AI models.
• Focus on two main aims:
  1. Create or improve generative AI models to perform complex mental health tasks safely and effectively.
  2. Produce evidence on how generative AI should collaborate with clinicians and end users.

Target Beneficiaries:

• Research teams with expertise in mental health, generative AI, clinical input, AI ethics, and lived experience of anxiety, depression, or psychosis.
• People affected by anxiety, depression, and psychosis through improved measurement and treatment.
• Multidisciplinary teams including early-career to established researchers.
• Teams must include lived experience experts as collaborators or coapplicants.

Application Process:

• Two-phase process:
  1. Accelerator stage (4 months, not a grant): Pilot studies, upskilling workshops, collaboration setup, and pilot funding.
  2. Funding call (up to £3 million for 2 years): Only accelerator participants can apply.
• Teams (2-8 applicants) apply, not individuals.
• Matchmaking service available to form teams.
• Lead applicant must have appropriate contract and management experience.
• Coapplicants can be from eligible organizations worldwide (except mainland China).
• Applications assessed on team expertise, potential impact, feasibility, and ethical considerations.
• Application deadlines: Accelerator applications open March 24, 2025; deadline May 28, 2025; funding call opens August 2025.

Key Features:

• Focus on anxiety, depression (including OCD, PTSD), and psychotic disorders (schizophrenia, bipolar, postpartum psychosis).
• Emphasis on ethical AI use, bias evaluation, transparency, and lived experience involvement.
• Support includes workshops, pilot funding, ethics guidance, and industry collaboration opportunities (e.g., Google).
• Research must not deploy AI solutions broadly during the award; only controlled experimental use allowed.
• Encourages innovative approaches to model development, evaluation, and human-AI collaboration.
• Administering organizations must comply with Wellcome grant conditions.

This program aims to advance mental health research by integrating cutting-edge generative AI with clinical and lived experience insights to improve outcomes for anxiety, depression, and psychosis sufferers in a responsible and impactful way.