AI Set to Speed Up HIV Vaccine Research and Trials, Focusing on Fair Access | IAS 2025

Artificial intelligence (AI) has the potential to transform HIV vaccine development by speeding up the vaccine design process and improving how data is analyzed and how clinical trials are run—especially in low- and middle-income countries—so real-world progress can keep pace with scientific breakthroughs, according to findings presented today at the IAS 2025 conference in Kigali, Rwanda.

During a session focused on AI’s influence in HIV vaccine research and development, researchers and innovators from leading organizations discussed how AI is poised to reshape vaccine research and clinical trial implementation for HIV.

Presenters included Jirair Ratevosian, M.P.H., a health policy expert at Duke University’s Global Health Institute; Rory Henderson, an associate professor of the Duke University School of Medicine and Shirley Collie, founder and CEO of clinical data management company BioInformatiCo. Each expert highlighted a different aspect of AI’s growing role in tackling how vaccines are designed, tested and made more available to those in low- and middle-income countries.

Ratevosian introduced how AI can help reach these more vulnerable communities who often face hurdles to testing and treatment.

Ratevosian, M.P.H.

Chatbots, for instance, can improve outreach and testing among high-risk populations, according to Ratevosian.

His team’s work has shown that using natural language processing (NLP) can mock up real conversations that feel personal and trustworthy.

“They reported over 90% of users engaged with the chatbot all the way to completion,” he said, adding results like that show AI has the potential to foster trust and provide accurate health information in ways that resonate with users utilizing the platform.

He added that 56% of users requested an HIV self-test kit, which is “impressive” given that many were young and first-time testers.

Additionally, AI can help connect vulnerable groups with healthcare providers. By looking at behavior data and giving tailored responses quickly, AI tools can reach more people in ways that fit their lifestyle, according to Ratevosian. This helps find HIV cases sooner and gives useful information to improve vaccine outreach plans.

Henderson then explained how AI is revolutionizing immunogen design and reshaping vaccine discovery strategies.

In terms of vaccine design and reshaping vaccine discovery strategies, outreach is one part of the puzzle. However, designing effective vaccines is another critical area where AI is showing promise.

According to Henderson, machine learning and protein language models are helping researchers understand HIV’s complex structure and how it can potentially evolve.

Henderson

“Modern AI tools provide a means to integrate complex data and recognize complex patterns, provide their training using the right data and (are) guided by the most informative features governing the underlying process that give rise to that data,” Henderson said. “If we train an AI algorithm to recognize the antigen features driving high precision antibody mutant selection, we can scale to 1000s of antibody targets on the hours today's timescale.”

He added that protein language models analyze amino acid sequences to generate detailed representations of HIV antibodies, helping researchers identify which mutations contribute to broad neutralization. By fine-tuning these models with diverse data, scientists can predict how antibodies will bind to the virus, improving their ability to design effective vaccines.

Henderson also noted the value of integrating data from multiple sources.

AI systems can combine insights from genomic sequencing, epidemiological surveillance and immunological research to refine vaccine targets. This could be especially useful for viruses such as HIV, which mutate rapidly and behave differently across regions. With AI, researchers can design vaccines that are more responsive to real-world viral diversity, rather than relying on static or narrow templates.

As far as clinical trials go, AI can reduce the research process where progress can typically slow down or get delayed due to challenges such as limited resources, complex procedures or inefficiencies, according to Collie.

Collie, who previously worked with Discovery Health in South Africa, focused on how AI can reduce delays and inefficiencies in clinical trials, especially in low- and middle-income countries where trial infrastructure is often fragmented.

Collie

“We work in complex, resource-constrained environments, and AI applied thoughtfully to systems can alleviate operational pain points,” Collie said. “It can help us unlock speed, scale and reliability. We need to modernize trial systems, not just discovery tools if we want to deliver results faster, particularly for diseases like HIV that demand urgency.”

She outlined how AI can automate the setup of clinical trial databases, review protocols for logical errors and even assist with regulatory compliance. These advances could allow more local institutions in Africa and elsewhere to run trials independently and at scale, without having to rely on expensive external partners.

In one example, she described a large HIV vaccine trial that generated dozens of protocol versions, memos and bulletins—an environment where AI-driven document management and large language models could dramatically reduce confusion and error.

Collie also stressed the importance of scaling these tools equitably.

Many early-stage biotech companies and academic institutions in the Global South lack centralized data coordinating centers, she said. AI can help level the playing field by providing standardized, automated solutions that don’t require massive up-front investment.

Taken together, these insights show that AI is not a silver bullet, but it is a powerful catalyst for faster

Between each point shared amongst the professionals today, one common theme was evident: AI will not solve every problem in HIV vaccine development, but it can remove critical barriers that currently slow progress.

From accelerating vaccine discovery to improving how trials are conducted and how communities are reached, AI offers tools to modernize systems that haven’t kept pace with scientific innovation.

However, they shared the concern against seeing AI as a magic wand that solves everything instantly.

Ethical concerns, data quality and algorithmic bias all remain challenges—especially when implementing tools in regions with limited digital infrastructure. Though, if used thoughtfully, AI can significantly increase the speed, equity and effectiveness of HIV vaccine research.