Description

This funding opportunity from the National Institutes of Health (NIH), specifically the National Institute of Mental Health (NIMH) and the National Institute on Drug Abuse (NIDA), aims to leverage cutting-edge multimodal artificial intelligence (AI) technologies to accelerate HIV clinical care. The initiative focuses on developing, adapting, and evaluating advanced multimodal AI models that integrate multiple data types (e.g., text, images, audio) from diverse sources (e.g., medical records, wearable devices, patient surveys) to improve HIV diagnosis, prevention, and treatment.

A key innovation encouraged by this program is the creation and integration of HIV-specific knowledge graphs to enhance the interpretability and explainability of AI model outputs. These knowledge graphs represent complex relationships among HIV symptoms, treatments, social determinants of health, and other relevant factors, enabling more precise and transparent AI-driven clinical decision support.

The program emphasizes a human-centered AI approach, requiring meaningful engagement with community members, patients, clinicians, and other stakeholders to ensure ethical, fair, and unbiased AI systems. Applications must address data privacy, security, transparency, and consent, and incorporate best practices for AI model development and deployment.

NIDA encourages applications that address HIV in the context of substance use disorders, including integrating substance use data into knowledge graphs and developing predictive models to improve clinical outcomes for people who use drugs.

Clinical trials are optional but, if proposed, must comply with NIMH's experimental therapeutics approach, including specifying intervention targets and mechanisms.