When Emergencies Outrun Official Alerts
John Mills, co-founder and CEO of Watch Duty, started the organization after repeatedly living through a familiar frustration: wildfires burning dangerously close, helicopters overhead, and no official alert arriving until far too late. Most wildfires turn deadly in their first hour, he notes, yet emergency alerts often lag behind the actual threat. The best real-time information, Mills found, was coming not from agencies but from individuals who spent their time monitoring radio traffic and sharing updates through large social media followings—some commanding audiences of hundreds of thousands.
Watch Duty grew out of that observation. What began as passionate individuals working independently is now a community-led organization run largely by volunteers. The service remains free for residents, with revenue coming from selling information to utilities and emergency services. The scale of the need became clear to Mills when tanker pilots flying multi-million-dollar aircraft thanked him for the intel Watch Duty provided—more than they had ever received before.
Data Science for Displaced Populations
USA for UNHCR's The Hive takes a different angle on emergency response. Seema Iyer, senior director of the data science innovation lab, oversees work that sits within UNHCR's broader mission of protecting forcibly displaced people worldwide. That population has more than doubled in the last decade, driven by crises in Syria, Afghanistan, Ukraine, and Venezuela.
The Hive was created ten years ago with a mandate to harness data science and technology both to raise awareness about refugees and to design practical solutions for their needs. One current project, run with UNHCR Kenya and supported by a Microsoft grant through their AI for Good Lab, uses drone imagery and machine learning to map camp infrastructure. Refugees arriving in western Kenya from different parts of Africa receive shelter materials, but UNHCR doesn't always know precisely where they settle. The machine learning model identifies where people have set up shelters and other infrastructure, supporting planning and resource management within the camp. The code and models were built to be open source from the start, with final outputs available on GitHub.
The Tension of Openness
Both organizations navigate a careful balance between openness and sustainability. Mills, a 30-year software veteran who has relied on open source throughout his career, describes himself as "torn internally." Watch Duty is partly open source and shares a lot of open data, but Mills is pragmatic about limits: other apps are attempting similar work and charging for it. Opening up everything would let them take the work Watch Duty has done, charge for it, and leave the organization without a future.
Iyer echoes that tension. Even nonprofits need a revenue model to cover costs. But she emphasizes that open source serves as a critical engagement tool. The Hive curates pathways for people curious about the mission to learn about the topic and apply their skills—potentially generating solutions the organization hasn't considered. A Microsoft-built model that identifies solar panels within a refugee camp, for example, might be adaptable to other contexts, saving others from starting from scratch.
That dissemination is, for Iyer, the signature of an innovation lab. Sharing what's learned enables others to build on it rather than duplicate it, and vice versa. It's also how the humanitarian sector can make use of skills that would otherwise be out of reach: technical expertise from civic technologists is often prohibitively expensive or simply inaccessible for humanitarian organizations. Curating those pathways of knowledge sharing is exactly what The Hive is trying to do.
Turning Participants Into Responders
For Mills, the most powerful outcome of community-driven emergency work is transformation. A small number of people can really change the world, and software enables that. He describes volunteers becoming first responders in their own right—firefighters meeting them and shaking their hands. "That's the beautiful part about what we can do with machines when they're used properly," he says. "They don't replace us, but they enable us."
