The following guest post was written by Yolanda Botti-Lodovico. Botti-Lodovico is the Storytelling and Advocacy Lead at the Patrick J. McGovern Foundation (PJMF), a 21st-century philanthropy advancing artificial intelligence and data solutions to create a thriving, equitable, and sustainable future for all.
An article published in Women’s International Studies Forum defined gender transformative approaches as those aiming to “reshape gender dynamics by redistributing resources, expectations, and responsibilities between women, men, and non-binary gender identities, often focusing on norms, power, and collective action.” In global health, such approaches go hand-in-hand with efforts to increase the agency and autonomy of all patients, regardless of their gender. But for the 25% of women surveyed by the UNFPA – many in sub-Saharan Africa – who are unable to make independent decisions about their health care, as well as marginalized women in high-income countries who report being unheard or discriminated against by medical providers, personal agency and autonomy remain largely out of reach.
AI in Gender-Transformative Care
AI solutions are already advancing gender-transformative care around the world. But health systems benefit most from new digital innovations when merged with robust, integrated infrastructure, people-centered practices, and commitments to equitable, quality care.
Topaz Mukulu, Strategy Analyst at the Patrick J. McGovern Foundation (PJMF), states that the most effective tools start from an understanding of which social determinants drive health outcomes. “We can then build AI applications that create pathways to greater agency, dignity, and personalized care for each patient,” she states. In line with the principles of gender-transformative impact, many of these tools are designed to confront deeply rooted norms in women’s health, expand women’s power across the continuum of care, and promote collective action through large-scale data analysis.
For example, Intelehealth created a free, open-source telemedicine platform, powered by an AI-enabled digital assistant, Ayu. It creates a direct line of communication between frontline community health workers and doctors (provider-to-provider), and between patients and doctors (direct-to-patient). Dr. Neha Verma, the CEO and co-founder, emphasized in a recent interview with PJMF how uniquely transformative such a platform can be for women in India, 90% of whom can’t make independent decisions about their health care and 50% of whom are unable to travel to a health facility without a companion. The team is refining their AI technology to collect and summarize information from Ayu, produce a likely diagnosis, and support clinical decision making in resource-constrained settings. At scale, such an intervention has the potential to not only challenge gender norms as more women seek care independently, but also increase women’s access to relevant, reliable, and timely information about their health.
Using deep learning techniques, AI is also supporting early detection of life-threatening diseases that disproportionately affect women. After learning that more than half of all breast cancer patients seeking care at a central hospital in Brazil had already advanced to a late stage of breast cancer, Instituto Protea used AI to create a risk model for earlier breast cancer detection based on mammogram screenings. In an overburdened and resource-constrained public health system – upon which 75% of Brazilians depend – early, AI-enabled diagnosis holds great potential for reducing overall costs of detection and treatment, expanding women’s power along their treatment journeys, and increasing quality of care and chances of survival.
For improved health equity at scale, AI is enabling greater collaboration and knowledge-sharing between stakeholders and researchers. Michelle Pomerantz, Board Liaison at PJMF, emphasizes that “we can’t make informed decisions about something we are failing to measure.” Accurate, robust, and diverse data sets can inform new AI tools for personalized and culturally responsive care, while uncovering and addressing systemic barriers to health equity.
Building robust, population-level data sets on maternal health is especially relevant in the U.S., where maternal mortality rates are both the highest of developed countries and rising. Non-Hispanic Black women suffer disproportionately, with a mortality risk almost triple that of non-Hispanic white and Hispanic women. In response, the Scripps Research Digital Trials Center developed the PowerMom platform to collect health information from diverse populations during pregnancy via wearable sensors and app-based surveys. Using privacy-preserving methods of data collection, the platform uncovers how systemic racism affects pregnancy and birth outcomes, which can then inform better practices and policies in reproductive health care. The team is simultaneously exploring LLM integration to produce personalized in-app health guidance.
Pillars of Success
While AI can open the doorway to gender-transformative health care, it’s insufficient on its own. The successful integration of AI in health care depends on four foundational pillars.
- Robust governance emphasizing data security and privacy
Dr. Verma noted that data privacy is paramount when integrating off-the-shelf LLMs with digital health applications for diagnosis and care. Without robust data security frameworks at an organizational level or via local policies and regulations, the patient could suffer errors in diagnosis, treatment, or other forms of harm.
In some contexts, a privacy breach could put both the patient and health care provider at risk. After the Dobbs v. Jackson Women’s Health Organization decision overturned Roe v. Wade, concerns abounded over the unauthorized use of AI to collect reproductive health information. Movements to criminalize those performing or seeking abortions in certain states “[have] undermined trust between patients and health care providers” and compromised the quality of care according to the Center for Democracy and Technology (CDT). Where federal standards for data privacy are lacking, the CDT notes the importance of shield laws and commitments from companies and providers to “minimize the collection, storage, and sharing of sensitive health data.”
- Buy-in from communities and stakeholders across all levels of health care
In addition to legal privacy protections, patients and providers often require guarantees that AI-assisted diagnosis and care decisions will be effective. Mukulu underscores the importance of “transparent communication on how AI systems work, including their limitations, safeguards, and the technology’s usefulness for the local context.”
She recommends engaging in a “stakeholder mapping exercise” to identify local and global champions and include them in the integration process. Close collaboration with community leaders and civil society can rebuild trust with marginalized groups previously excluded from or harmed by AI’s use in the medical field.
- Enabling ecosystems
AI solutions operate best within a health ecosystem well-equipped to meet local health needs and supported by an interoperable digital infrastructure. But talent gaps, resource constraints, connectivity issues, incomplete data stores, and “AI pilotitis”, according to Mukulu, prevent many health systems from integrating AI safely and sustainably beyond the project period.
To avoid AI pilotitis and support broader health ecosystem strengthening, Intelehealth provides comprehensive support for implementation of its telemedicine platform and capacity building for providers across both the NGO and public sectors. Their digital assistant, Ayu, is also designed to operate in remote and disconnected regions, with offline data collection capabilities and easy-to-use features for patients with limited digital literacy.
- Broad-based funding for comprehensive and equitable research
The benefits of AI in health and beyond are predominantly concentrated in the Global North. This is in part due to structural barriers in the Global Majority, including limited technical infrastructure and capacity, computing costs, as well as availability and compatibility of data sets. Additional investment in research and development efforts can help create a pathway to transformative solutions trained on Global Majority data sets, such as Instituto Protea’s breast cancer AI technology. Moreover, supporting relevant health innovations for communities anddiseases that suffer from underinvestment is paramount to gender-transformative health equity on a global scale.
Charting a Path Forward
Health care for women becomes transformative when it’s combined with comprehensive systems change – starting with deeply embedded norms around women’s agency and autonomy. Emerging AI applications are already demonstrating significant potential, empowering women to define and own their health journeys. For impact at scale, commitments to local health systems strengthening will be key to bolstering efforts from innovators and researchers building new health solutions, as well as advocates working to uphold patient rights and dignity for sustained health equity worldwide.
