Content provided by the The Lancet (https://www.thelancet.com) Note: Content, including the headline, may have been edited for style and length.
Artificial intelligence is now influencing nearly every part of modern life, changing how people work, communicate, and manage their health. HIV care is no exception. AI is beginning to shape research, clinical services, and public-health strategy, with growing interest in how these tools can improve the lives of people living with HIV, support health-care providers, and strengthen the wider response to the epidemic. But for that promise to be realized, HIV experts, communities, and technology developers must work together to ensure AI is introduced safely, responsibly, and fairly.
Interest in this area is rising quickly. The past two years have seen a sharp increase in work exploring how AI can support HIV prevention and care, especially as health systems face financial pressure and shifting priorities. Across Africa, decades of HIV research, digital health investments, and electronic records have created large volumes of clinical, behavioural, and epidemiological data that could support useful AI applications. In practice, this could mean improving service delivery, helping patients navigate prevention tools, and enabling more responsive care through better use of data.
Yet the excitement around AI should not obscure its risks. Recent controversies involving inaccurate AI-generated health information have shown how dangerous it can be to treat automated outputs as reliable medical authority. In HIV care, those risks may be even greater. Systems trained largely on data from well-resourced settings may not perform well in contexts marked by fragmented services, limited laboratory access, or the exclusion of key populations from available datasets. Without careful design, AI may reinforce the very inequities it is meant to address.
Questions of privacy and trust are equally serious. In many settings, HIV-related stigma remains strong, and in some places criminalisation still shapes the experience of those affected. Under such conditions, the handling of personal health data cannot be treated lightly. If AI systems are opaque, poorly governed, or vulnerable to misuse, they may expose sensitive information or increase fears of surveillance. That is why data protection, transparency, and local accountability must be central to any AI strategy in HIV programmes.
The HIV response has long been shaped by collaboration among patients, affected communities, clinicians, researchers, and programme leaders. That tradition should guide the next phase of AI adoption. Communities should not simply be consulted after systems are built; they should help shape how these tools are designed, evaluated, and overseen from the beginning. Their participation can challenge hidden assumptions in data, improve relevance across social and cultural contexts, and ensure technology serves both individual dignity and public-health goals.
AI may become an important part of the future of HIV care, but its success will not be determined by technical sophistication alone. It will depend on the ethical decisions surrounding its use: who builds it, whose data informs it, who benefits from it, and who is protected from harm. The goal should never be to replace clinicians or weaken human relationships in care. It should be to equip health workers and public-health systems with better tools while preserving trust, equity, and person-centered care. If guided well, AI could help close long-standing gaps in HIV services. If introduced carelessly, it could become one more barrier in a field that has already fought too hard against exclusion and inequality.
Content provided by The Lancet (https://www.thelancet.com) Note: Content, including the headline, may have been edited for style and length.
