Research Summary

AI Chatbot Use Associated With Higher PrEP Initiation Among US Adults

Anthony Calabro, MA

Key Highlights

  • Adults who interacted with the AI chatbot were more than 3 times as likely to initiate PrEP as non-users.
  • PrEP follow-up resulting in a prescription refill occurred among 57.0% of AI users versus 32.4% of non-users.
  • Appointment attendance was higher among AI users than non-users—66.1% versus 53.9%.
  • Because the study was observational, the findings establish associations rather than causal effects.

Interaction with an artificial intelligence (AI)–powered chatbot was associated with higher HIV pre-exposure prophylaxis (PrEP) initiation, follow-up, and appointment attendance in a large US cohort, according to findings published in npj Digital Medicine. Among adults without previous PrEP use, initiation rates were 19.3% among chatbot users and 6.1% among non-users.

Researchers conducted a retrospective cohort study across AIDS Healthcare Foundation Wellness clinics after a conversational AI messenger was launched in January 2024. The analysis included 155,217 HIV-negative adults with at least 1 clinic visit between January 1, 2024, and April 29, 2025. Of these participants, 36,394 interacted with the chatbot and 118,823 did not.

The chatbot delivered education and reminders concerning oral PrEP, appointment scheduling, and sexual health but was not designed to provide individualized clinical guidance. The primary outcomes were new PrEP initiation, retention defined as at least 1 follow-up visit resulting in a prescription refill, and attendance at a scheduled PrEP appointment after the introduction of AI-generated reminders.

Study Findings

New PrEP prescriptions were documented for 7012 of 36,394 AI users (19.3%) and 7268 of 118,823 non-users (6.1%). AI users were more than 3 times as likely to initiate PrEP (risk ratio [RR], 3.15; 95% CI, 3.00-3.27; P<.0001).

Among 13,831 individuals who initiated PrEP in 2024, 3841 of 6623 AI users (57.0%) returned for at least 1 follow-up visit resulting in a prescription refill, compared with 2340 of 7208 non-users (32.4%; RR, 1.78; P<.0001). Among 15,515 appointments evaluated after the reminder feature launched, attendance was 66.1% among AI users and 53.9% among non-users (P<.0001). Among participants who returned for follow-up, AI users returned a mean of 14.0 days sooner than non-users.

Clinical Implications

According to the study authors, AI-supported communication may improve early PrEP uptake and follow-up in real-world HIV prevention settings and complement clinician- and community-led efforts to reduce barriers to preventive care.

The authors cautioned that participants were not randomized and that people who chose to interact with the chatbot may have been more motivated or more likely to seek care. The analysis also lacked multivariable adjustment and did not measure medication levels, long-term PrEP continuation, or HIV incidence.

Expert Commentary

“Our results indicate that such tools may help improve early uptake and follow-up in real-world prevention settings,” the researchers concluded.

Reference
Narayan A, Blasingame M, Warmuth S, et al. AI-augmented communication improves HIV PrEP initiation and persistence in populations disproportionately impacted by HIV. NPJ Digit Med. 2026;9:411. doi:10.1038/s41746-026-02519-3