Transcript: Episode 5

Welcome to Impact Signals. Your daily briefing on how AI is reshaping disaster response, humanitarian aid, and social good. I'm Charlie here with Sarah. It's February 17th, 2026, and there is a lot happening. There really is. The India AI Impact Summit is in full swing in New Delhi, and today saw some major announcements. Let's start with the big one. India just became the first country in Southeast Asia to adopt a comprehensive national AI strategy for healthcare. That's right. Health Minister JP Nada launched two frameworks, Sahi, which stands for Strategy for Artificial Intelligence in Healthcare for India, and both a benchmarking open data platform for health AI. The WHO called it a landmark. And this isn't theoretical. India is already using AI in its public health system for a TB risk prediction, diabetic retinopathy screening, and telemedicine. What both adds is a way to benchmark these AI tools. Make sure they're actually safe and effective. Meanwhile, at the same summit, the UN's Disaster Risk Reduction Office dropped a reality check. UNDRR Suggit Mahanti said that AI disaster tools are essentially useless without legal frameworks backing them up. This is such an important point. Imagine you're a district magistrate, and an AI model tells you a flood is coming. Evacuate now. But there's no legal basis for that decision. You can't act on an algorithm's prediction. Let's talk about children. UNICEF released version 3.0 of their AI and children's guidance. This one has teeth. It really does. They're calling on governments to mandate child rights impact assessments, CRIAs, for any AI system deployed in public services that touches children's lives, 10 core requirements, including safety by design, strict bans on biometric data harvesting from children and refugee settings, and algorithmic audits for fairness across linguistic groups. UNICEF India's representative said AI must be safe by default. No more move fast and break things when we're talking about kids. And think about where AI intersects with children's lives. Education platforms, social services, refugee registration systems, the stakes are enormous. Getting this right early matters more than moving fast. And it's not a pilot anymore. This is a fully integrated municipal emergency service. Stockholm is operationalizing it. Now, the humanitarian funding crisis. We need to talk about this because it's the back of our backdrop to everything else we're covering. That stretches every dollar further isn't just nice to have. It's existential for these programs. Their follow the forecast system also uses predictive data for anticipatory cash transfers, getting money to people before disasters hit, not after. Quick look at active disasters. A magnitude 6.0 earthquake hit Russia on February 16th. The Solomon Islands had a 5.7 the same day. Fled alerts are active across the US, Peru, Turkey, Afghanistan, Ukraine, and Montenegro. Sudan's conflict is approaching its fourth year, and Yemen is tracking rapid displacement. And the India AI Summit runs through February 20th. We'll be covering it all week. If you're in the AI for good space, this is the event to watch. February 17th. Subscribe wherever you get your podcasts. Find us at impactsignals.ai. I'm Charlie. And I'm Sarah. See you tomorrow.