Writing

What we've learned building AI for young people, written down so others can use it.

2026 · Girl Effect · Written by me Building carefully: how we're researching AI and youth mental health Young people are already turning to AI with their worries. This paper sets out why we chose to study it rather than stay away, how we're designing a bounded, single-session GenAI support for young people in South Africa with safety built in from the start, and the questions we still can't answer. Read the paper → 2026 · World Bank Expanding access to mental health care through digital technologies A compendium of lessons from low-resource settings. Girl Effect is a case study: our chatbots in South Africa, India and Kenya, and how a high-risk conversation hands off to a trained person. Read the compendium → 2025 · Yale School of Public Health Big Sis: choosing a path to scale a mission-driven AI chatbot A teaching case on Big Sis, our AI chatbot for young women in South Africa. It follows the choices we faced as generative AI moved fast and donor priorities shifted: when to scale, how to stay safe, and what to protect along the way. Read the case → 2025 · Stanford Center for Digital Health Generative AI for Health in Low and Middle Income Countries A white paper on how large language models can be designed and governed for health where resources are tight. Our chatbots feature as a case study of what it takes to do this responsibly. Read the paper → 2025 · Girl Effect Building with GenAI: our journey to smarter, safer health chatbots Field notes from moving our chatbots onto generative AI. What worked with low-resource languages, why human safeguards still matter, and how we govern the data young people share with us. Read the field notes → 2025 · Girl Effect Ethical AI guidelines for social and behaviour change chatbots The rules we hold our chatbots to on safety, privacy and moderation. Written in plain language and shared openly, so other teams building chatbots for young people can use them and make them better. Read the guidelines → 2023 · Girl Effect Artificial intelligence and machine learning vision Where we said we were going with AI, before most of the sector was talking about it. It sets out how machine learning could help us reach more young people with the right information, and the guardrails that had to come with it. Read the vision → 2020 · UNICEF When chatbots answer their private questions A learning brief on safeguarding in chatbots for young people. It features our early keyword triggers, which spot when someone shares something serious and route them to trained staff. Read the brief →

Speaking

From the Skoll World Forum and the AI for Good Summit in Geneva to Yale and a UNGA side event, I talk about building AI that young people can actually trust.

2026UNGA side event hosted by Girl Effect, New YorkWho shapes what comes next in AI? What if girls did.
2026Skoll World Forum, OxfordDesigning AI for Transformative Change · Watch →
2025AI for Good Global Summit, GenevaBridging frontier innovation with non-profit practice
2025Yale Innovation SummitGlobal Health at a Crossroads: Innovation and Investment in a Changing Era
2025Global Digital Health ForumFoundations and Futures: Reimagining Public Health Intelligence in the AI Era

Invite me to speak →