Transcript: Episode 9

Welcome to Impact Signals, Social Impact at the Scale of AI. I'm Charlie. And I'm Sarah. Saturday edition today, and it's a big one. The India AI Impact Summit just wrapped and the outcomes are significant. Let's start there. 88 countries just signed the Delhi Declaration, the most broadly endorsed global AI governance statement we've ever seen. Both the US and China signed on, which almost never happens. That's the headline worth watching. Getting geopolitical rivals to co-sign anything on AI is extraordinary. The declaration covers seven pillars and one maps directly to our space, resilience, innovation, and efficiency. That's disaster management language embedded in a global AI framework. For humanitarian organizations, this creates a multilateral reference point you can actually cite and policy advocacy. And Geneva hosts the 2027 follow-up. So this isn't a one-off. It's building a governance pipeline. The government innovation fund is the one to watch. The call is open to partnerships that transform essential services to complex societal challenges. Disaster response agencies, humanitarian organizations, this is your application window. And the Climate Technology Center with India's principal scientific advisor signals serious investment in AI-powered climate adaptation. Which disaster risk is climate-driven? These tools could feed directly into preparedness systems. Meanwhile, at the same summit, UN Secretary General Guterres launched a 40-member independent international scientific panel on AI. He was blunt, said AI innovation is moving at the speed of light and that human oversight must be a technical reality, not a slogan. Back in the U.S., a very different story. FEMA's Disaster Relief Fund has dropped from $30 billion in December to $9.6 billion. At least $17 billion is held up for review. And a new DHS bulletin requires all FEMA travel to be pre-approved commercial government shutdown. In the Indian Ocean, tropical cyclone gazani hit Madagascar hard. The UN launched a $67.8 million flash appeal. to dead, 382,000 people needing urgent aid. The scale of that flash appeal tells you the infrastructure damage is severe. 25 districts across five regions affected. What's notable from a technology standpoint, UNOSAT deployed satellite-based damage mapping and map action published population level impact maps by district. This is remote sensing-driven needs assessment operating at scale. Exactly the kind of AI-assisted humanitarian response we track. China's 100 million yuan pledge and France sending rescue teams from nearby reunion show the international coordination mechanism activating. This is a recovery model worth studying. You've got the state government, opposition parties, and civil society organizations all running parallel reconstruction programs. The timeline pressure is real. They need to finish before monsoon season. That disaster cycle reality driving recovery urgency is something every practitioner understands. And the government extending loan moratoriums shows the financial support layer that makes physical reconstruction viable. And in Southeast Asia, Tokyo-based Spec T is expanding its AI disaster prevention platform to Vietnam and Indonesia, backed by Japanese government funding. This is interesting technology transfer. Spec T Pro uses real-time social media and sensor data for disaster detection and early warning. They've already launched in the Philippines, so there's operational proof of concept. Japan funding AI disaster resilience tools across Southeast Asia could become a model for how developed nations support climate vulnerable regions. It's donor-backed transfer with built-in sustainability. The tools are designed for the local context, not retrofitted from Western deployments. Finally, the U.S. State Department is proposing a new Bureau of Disaster and Humanitarian Response, separating disaster relief from migration programs. Big structural change happening there. Coming alongside USAID layoffs and program absorption by state, this could reshape how the largest bilateral donor in the world delivers humanitarian assistance. Separating disaster response for migration could streamline emergency funding, or it could create new bureaucratic silos. Organizations receiving U.S. humanitarian funding need to track this restructuring closely. The continuity risk for existing programs is real. Coming up, NVIDIA GTC in March with workshops on edge AI for field deployments, and the UN Global Dialogue on AI Governance in Geneva this May. We'll be watching both. Big quarter ahead for AI governance and humanitarian tech. That's your Impact Signals Weekend Briefing. Subscribe wherever you listen, follow us at ImpactSignalAI on X, and visit impactsignals.ai. Until next time, stay ready.