ICTworks

Technology is transforming access and delivery of sexual and reproductive health (SRH) information and services, and artificial intelligence (AI) is one of the latest additions to this technological advancements. From fertility tracking to maternal health, AI is expanding opportunities to potentially more personalized and timely care.

How AI is supporting sexual and reproductive health and rights

AI is already influencing the full spectrum of SRH services, ranging from contraception, maternal health, infertility and fertility care, sexually transmitted infections, cervical cancer, and abortion.

Based on a technical brief published by WHO, these are seven key leading areas in which AI is being applied in SRH. WHO is now in the process to compile the evidence and evaluate the cost-benefit, acceptability, feasibility and equity implications of these emerging uses.

1. Health information, education and promotion

AI models can be leveraged to develop interventions for health education and to promote health behaviours. For example, virtual conversational agents or chatbots can provide information in what may be perceived as a more anonymous and non-judgmental manner compared with personal interactions.

These tools, often used directly by individuals, are gaining traction for their potential to overcome access barriers to traditionally stigmatized and sensitive areas of health care, such as sexual health, contraception and STIs.

2. Screening and diagnosis

AI can analyse large volumes of health data—from electronic medical records and medical images to lab results and clinical notes—to identify trends, patterns, and risk factors.

This includes interpreting imaging data to detect abnormalities such as cervical pre-cancer lesions, or predicting which pregnant women may be at risk of complications like postpartum haemorrhage, pre-eclampsia, gestational diabetes, or preterm labour.

AI algorithms can also be integrated into medical devices, such as ultrasound machines, to detect fetal distress and support task sharing, especially in settings with limited access to specialists like radiologists or sonographers.

3. Treatment and care management

AI’s predictive capabilities can also support personalized treatment plans. For instance:

• AI algorithms help tailor in vitro fertilization (IVF) in fertility care, by selecting sperm, oocytes, and embryos, and predicting IVF outcomes.
• AI can combine different data sources to predict the need for insulin in pregnant women with gestational diabetes.
• AI is being used to optimize antiretroviral therapy dosing for people living with HIV, helping to improve care and reduce side effects.

4. Personal health monitoring

A common use of AI in SRH that has grabbed media attention is personal health monitoring through wearable devices and mobile phone apps. This may include personalized fertility trackers using AI to predict ovulation and fertile windows, and AI-based tools for symptom management, such as during pregnancy and menopause.

5. Understanding health trends

AI can analyse large-scale data to monitor public health trends and identify emerging issues. This includes detecting links between different data sets—such as environmental exposures and maternal health, analysing social media interactions for insights on public perceptions on topics such as STIs, abortion, HPV vaccines, and monitoring global health data to track trends in contraceptive use.

6. Health system management

The predictive modelling functions of AI can enable forecasting of needs and assist with targeting interventions. Examples may include predicting patient inflow, emergency transport needs, and commodity stock-outs to optimize inventory management for medicines and other supplies.

7. Clinical research and drug discovery

AI can assist researchers and clinicians in analysing complex data sets to accelerate clinical research and drug discovery. Examples include using:

• AI algorithms to model molecular and genomic data to predict outcomes related to new therapeutics and drug resistance, particularly for HIV,
• Machine learning models to identify genetic markers associated with reproductive health conditions.
• AI for the development for treatments related to pregnancy and infertility, among other conditions

Navigating challenges with care

As promising as AI is, it must be used responsibly to avoid unintentional harm, particularly on sexual and reproductive health and rights (SRHR), which deal with controversial health topics and underscore the critical need for data governance. Key considerations include:

1. Data governance and bodily autonomy

With the rise of personal health monitoring—such as fertility tracking—and AI tools for health information, education, and promotion, more SRH data are being generated outside traditional medical settings.

As a result, these data may not be protected under existing regulations, such as the Health Insurance Portability and Accountability Act (HIPAA) in the United States. This lack of protection can undermine individuals’ rights to health services and bodily autonomy—for example, when menstrual health data are collected and shared with third parties for targeted advertising or, in some cases, used for surveillance of pregnancy terminations.

2. Misinformation and targeted disinformation

A major challenge in using AI for sexual and reproductive health and rights (SRHR) is the risk of misinformation. Many AI models are trained on large datasets from the internet and social media, which often contain low-quality information and various biases.

In addition to unintentional misinformation, there is also a risk of deliberate manipulation or disinformation, in which content may be designed to discourage individuals from seeking SRH services.

3. Context and cultural awareness

The foundations of text- or speech-based AI technologies, such as LLMs, rely heavily on language, which requires a localized or culturally nuanced approach in the field of SRHR.

For example, discussions related to contraception can be influenced by personal beliefs, psychology, culture and socioeconomic factors. As most AI systems and tools are trained on a minority of well-resourced languages, the limited representation of other languages can create potential misalignment with local context and limited user engagement.

Moving forward responsibly

While we all grapple with the common challenges presented by AI, the unique norms and power dynamics within sexual and reproductive health and rights (SRHR) amplify certain risks and concerns.

AI systems and tools are not inherently harmful; however, the specific ways AI is applied—and the risks involved—are shaped by the policies, social contexts, and power structures surrounding SRHR.

To fully realize AI’s benefits in SRHR, we must carefully address the following key challenges:

1. Revisit data protection regulations and redress mechanisms

As data is increasingly collected and commodified for AI use, data protection laws must be strengthened to prevent and address potential breaches. This includes clear limits on data use, sharing, and repurposing.

2. Fight misinformation and targeted disinformation

Implement community-led, open-source fact-checking programs and develop transparent interfaces that clarify AI-generated recommendations. Partner with local health workers and community leaders to disseminate accurate, culturally relevant information that supports marginalized or vulnerable groups. Establish certification standards to verify that chatbots and conversational agents provide fact-checked content.

3. Promote inclusivity and data diversity

Encourage diversification of data sets used to train AI algorithms, ensuring representation across socioeconomic, educational, and cultural backgrounds. Build local and national capacity to develop relevant data sets, case repositories, and language resources to better contextualize SRHR needs.

4. Establish collaborative oversight mechanisms

Use approaches like “human-in-the-loop” to actively detect and mitigate biases and inaccuracies. Strengthen the integration of human rights and SRHR-specific risk considerations within broader AI risk management frameworks.

Furthermore, while AI systems classified as Software as a Medical Device (SaMD), such as AI-assisted ultrasounds require regulatory approval, there is also a need for adaptable oversight pathways to monitor AI applications integrated into consumer-facing apps, especially as more individuals use these tools to access SRH information and services.

Tigest Tamrat is a Scientist at World Health Organization