How AI is Revolutionizing HIV Prevention and Care

Have you ever used a smart phone app or talked to a digital helper online? These tools use Artificial Intelligence (AI), which means smart computers that can learn and solve problems. Now, experts are finding exciting new ways to use AI to improve the health and lives of people affected by HIV. AI is becoming a powerful tool that can help people get accurate information, find support, and reduce the costs of health campaigns.¹

What Is AI and Why Does It Matter for HIV Care?

AI is essentially the ability to turn huge amounts of different information into useful ideas that lead to better decisions and results. In healthcare, AI can look at massive health data volumes much faster than a human.¹ This capability is especially important because it can increase efficiency and accuracy when funding for health services is being cut.²

By spotting patterns, AI can help predict problems, such as a possible disease outbreak.³ It can also help guess when a person might stop coming to their doctor appointments. This allows health workers to respond earlier to prevent gaps in service. AI makes it possible to offer more personalized care that is scaled up in a way that is easy to sustain.⁴

AI Finds Better Health Messages Online

One way AI is making a difference is by improving how public health departments share important information. Traditional campaigns used to be very expensive and often selected messages without a clear system. Now, researchers are using machine learning, a type of AI, to find helpful HIV prevention messages that already exist on social media platforms.²

This process creates what are called "living health promotion campaigns".¹ The AI model sorts through thousands of messages to find those that are "actionable." An actionable message is one that encourages a user to take a specific step, such as getting tested for HIV or starting PrEP, or shows where to find a testing center.¹

In a recent study, the messages chosen by the AI model were seen by gay and bisexual men as significantly more effective, more accurate, and more relevant to their lives than messages that were not selected by AI.² The men also reported that they would be more likely to share these AI-selected messages. Furthermore, public health groups in the United States selected and posted the AI-recommended messages about seven times more often than other messages. This shows that using AI can help public health departments promote health with fewer resources.⁴

Chatbots Provide Secret Support

Another exciting use of AI is through chatbots. These are digital assistants that can support clients and answer their questions, often reducing the pressure on busy healthcare providers. Chatbots can fill waiting time at a clinic and help clients prepare for their consultation.⁴

AI for Young Women: The Aimee Chatbot

Aimee is an AI chatbot specifically designed for adolescent girls and young women in South Africa, aged 16 to 24. Aimee is built to be like a trusted friend, ready to chat about sensitive topics like PrEP, relationships, and HIV, or even worries about pregnancy and gender-based violence.⁴

Aimee works hard to build trust with users by using empathic responses and listening closely to the user’s mood. When users feel safe, they start to talk about deeper concerns that they might hide at first. Research shows that users talk to Aimee about relationship issues and gender-based violence much more frequently than they talk to nurses. For clients who chat with Aimee, about one-fourth have taken action on a service the bot offered, such as seeking out HIV testing or starting contraception.⁴

Helping with Tough Feelings: The MARVIN Bot

MARVIN is another chatbot, used in Canada, that helps people with self-managing their HIV.⁴ Developers have adjusted MARVIN to make sure it responds correctly to messages that show negative feelings, especially those related to self-harm or depression. The AI was trained to recognize very negative language so that it could appropriately offer emergency contact information for users who expressed thoughts of self-harm. In the future, MARVIN will be trained to detect more signs of psychological distress, such as anxiety.⁴

Making Sure AI Is Fair

While AI holds great promise, experts warn that it must be designed carefully. If the data used to train the AI systems is biased, the AI could end up repeating and speeding up the same inequalities already seen in care.¹

It is important to include the real-life experiences of all groups of people, especially those who are often ignored or made invisible by traditional data systems.⁴ If developers do not include this "lived experience," the AI may create "blind spots" and fail to help the people who need support the most. To truly improve care, AI must be a tool that listens and is built with trust and fairness in mind.⁴

Conclusion

Artificial Intelligence is already changing HIV prevention and care by creating targeted health messages and offering private, supportive digital companions.⁴ AI tools like Aimee and MARVIN show how smart technology can build trust and encourage people to take control of their health. As these technologies advance, it will be critical for developers to focus on creating systems that are fair and inclusive for everyone.⁴