An AI-based behavioral intervention to improve pre-exposure prophylaxis for HIV prevention - HIV continues to disproportionately affect men who have SEX with men (MSM), particularly those living in…

The National Institute of Mental Health awarded The Miriam Hospital a $705,322 Project Grant (CFDA 93.242, Mental Health Research Grants) on September 16, 2025, to develop and evaluate an artificial intelligence-based behavioral intervention designed to improve HIV pre-exposure prophylaxis (PrEP) uptake among men who have sex with men (MSM) in Ending the HIV Epidemic (EHE) jurisdictions. The primary deliverable is CHIA (Chatbot for HIV Prevention and Action), a generative AI-based chatbot tailored with motivational interviewing (MI) techniques integrated with the transtheoretical model (TTM) to assess HIV risk, evaluate readiness for change, deliver stage-appropriate education, and motivate PrEP initiation. The project will address the labor-intensive limitations of traditional MI-based interventions by leveraging artificial intelligence to scale accessibility and reach broader populations simultaneously. The three-year project (concluding September 15, 2028) includes an open pilot trial with 15 participants to evaluate CHIA's accuracy and participant satisfaction, followed by a pilot randomized controlled trial to assess acceptability, feasibility, and preliminary efficacy.

Sub-awardees include the University of Florida and Brown University, both contracted to contribute specialized research expertise to support the development, refinement, and evaluation phases of the CHIA intervention. These collaborative partnerships extend the research capabilities and scientific rigor of the primary investigative team, enabling multi-institutional expertise in AI-driven health interventions and behavioral sciences research focused on HIV prevention.

An AI-based behavioral intervention to improve pre-exposure prophylaxis for HIV prevention - HIV continues to disproportionately affect men who have SEX with men (MSM), particularly those living in the ending the HIV epidemic (EHE) jurisdictions. Despite the proven effectiveness of HIV pre-exposure prophylaxis (prep) in preventing HIV infections, its uptake among MSM is notably inadequate, particularly among those living in EHE jurisdictions. Motivational interviewing (MI) has shown effectiveness in encouraging at-risk MSM to adopt prep; integrating the transtheoretical model (TTM) into MI enables the assessment of an individual's readiness for change and the delivery of tailored MI sessions. However, its labor-intensive nature, requiring multiple sessions, poses a significant challenge. The potential solution lies in harnessing generative artificial intelligence (AI) to streamline and make MI more accessible. This pilot study aims to leverage AI technology to overcome the labor-intensive aspects of MI, potentially reaching a broader audience simultaneously. In the study, we aim to tailor, refine, and test an AI-based chatbot called chia (chatbot for HIV prevention and action) with MI techniques to promote HIV prep among MSM living in the EHE jurisdictions. The chia will be tailored and fine-tuned based on the latest chat generative pre-trained transformer (chatgpt) model to employ MI techniques to assess HIV risk and TTM stage, provide stage appropriate education, and motivate individuals to initiate prep (specific aim [SA] 1). In an open pilot trial involving 15 participants aged 18 years or older recruited via various social media platforms, we will rigorously evaluate the chia's accuracy, participant satisfaction, and responsiveness to individuals' needs with refinement made as indicated in SA2. The refined chia will subsequently undergo evaluation in a pilot randomized controlled trial (SA3) to assess its acceptability, feasibility, and preliminary efficacy in promoting prep uptake. We will randomly assign participants to the intervention (at least 60 minutes of engagement with chia in two or more sessions within two weeks of enrollment, n=50) or the control arms (standard of care by research staff: risk assessment, prep education, risk reduction counseling, and prep referral if interested, n=50) with a 1:1 ratio. Primary outcomes include prescription receipt, pharmacy filling, and drug concentration levels, providing definitive markers of prep uptake. The proposed pilot study has the potential to leverage the power of emerging AI technologies to incorporate MI into HIV prevention and prep promotion. If successful, the proposed pilot study will help facilitate AI in the field of HIV prevention as well as reduce HIV incidence among MSM.