When the warnings don’t work, but the Wi-Fi does: How communities are rewriting the future of climate response

This article is based on an Ignite Stage presentation at the 2025 Global Platform for Disaster Risk Reduction. The Ignite Stage offers fast-paced, impactful talks from diverse voices working at the forefront of disaster risk reduction.

While global efforts have poured resources into improving forecasting and modelling, a critical question remains: What use is a forecast if it is not trusted or relevant?

In 2023 during the wildfires in Maui, Hawai'i, sirens were never activated, and many residents did not know what the alerts meant or how to respond. During the 2021 eruption of Mount Semeru in Indonesia, the early warning system failed because it applied a US-based framework not adapted to local conditions; no warnings were issued for the rainfall-induced explosion.

The shortfall of early warning systems is neither new nor uncommon. Even the most sophisticated forecasts cannot account for what happens when hazards meet the messy dynamics of everyday life: aging infrastructure, densely populated urban areas and varying levels of trust in institutions. When critical information does not arrive in time, or is unintelligible, irrelevant or mistrusted, the consequences can be severe.

And yet, in the face of these shortcomings, people do not wait. Communities share information, organize evacuations and check in on neighbours using social media and messaging apps. These deeply responsive networks are rooted in trust, proximity and a shared understanding of what is at stake.

However in moments of crisis the sheer volume and urgency of information can quickly become overwhelming. Critical updates get lost in the noise, misinformation spreads, and unverified reports hinder coordinated response. While grass-roots efforts fill critical gaps, they are also often unsupported and unrecognized by formal systems, limiting their impact.

What if the same platforms people already use every day could serve as a coordinated, structured, trustworthy disaster information system?

Yayasan Peta Bencana (Disaster Map Foundation) has developed an open-source, scalable, disaster information system that transforms social media into a structured emergency coordination network. Our software plugs directly into existing social media channels and listens for disaster-related keywords, such as "flood", "fire", "earthquake" etc. When such a post is detected, an AI chatbot automatically replies with a simple prompt: Are you really in a disaster? This prompt invites residents to share verified information through a short survey that asks for the location, hazard severity, and, if possible, a photo and description. Verified reports are immediately displayed on live, real-time map (PetaBencana.id in Indonesia and MapaKalamidad.ph in the Philippines), freely accessible for all residents, first responders and government agencies.

Because social media posts are already geotagged and timestamped, this interaction unlocks a rich, stream of data. By structuring, sorting, and verifying this information in real-time, the system cuts through the noise and transforms everyday posts into timely, actionable insights essential for emergency response. No new app is required. The system meets people where they already are - on WhatsApp, Twitter, Facebook, etc. - reducing friction, increasing participation, and turning everyday digital behavior into life-saving infrastructure.

A report about a flood might be read by a hydrologist, a disaster agency, a taxi driver and a concerned neighbour, all using it to make different but interdependent decisions. The system's value lies in enabling those decisions to happen with shared reference points; distributed intelligence for collective action. In 2024 alone, the system was used by over 200 million residents and over 900 humanitarian agencies to guide response efforts. It is also integrated into local and national government systems, recognized as the fastest source of verified disaster information available.

According to one emergency manager in Sumatra, "Delays in emergency response due to bureaucracy have been completely eliminated by PetaBencana.id."

During and after a disaster, people make millions of crucial decisions: Where should I go? Is this road passable? Can I reach my family? These decisions are deeply local, context-dependent, and time-sensitive. And yet, most disaster management systems are designed for centralized decision-making. By making real-time disaster information accessible to everyone through their phones, PetaBencana.id democratizes the tools of response and shifts towards a model where everyone can see, contribute, and act.

Crucially, the system operates by recognizing local knowledge as critical infrastructure in its own right. High-resolution data is not the same as high-context data: local observations capture what satellites and algorithms miss, like blocked evacuation roads and stranded families. The Indonesian Meteorological Agency and even NASA now use PetaBencana reports to validate satellite models and improve flood prediction. These collaborations underscore how frontline insights fill the gap of hyper-local contextual data that are critical to strengthen global modelling systems.

Working alongside communities, we have found that the challenge is less about missing data and more about unrecognized knowledge. Instead of one-way streams of information flowing from "experts" to the public, warning systems need to be dynamic, networked infrastructures that actively leverage the rich insights of situational knowledge held by communities on the ground. Such systems of participation must be cultivated through trust, co-design and co-ownership, with community members invited to shape how data is gathered, shared, acted upon, and governed.

First launched in Indonesia in 2017, Disaster Map Foundation's software has scaled regionally through a model of South-South collaboration and open technology transfer. The expansion has prioritized local adaptation and co-design, ensuring each implementation reflects the cultural, institutional and geographic realities of the places it serves. Pilots are underway in India, Pakistan, Panama, Viet Nam and beyond. Mutual learning, where experiences in Jakarta inform deployments in Manila, which in turn influence pilots in Hanoi or Karachi, represents a powerful alternative to top-down tech interventions.

Looking more broadly, climate datasets must interrogate which knowledges are being amplified, and which are excluded. Whose experiences shape the narratives these systems reproduce? Without interrogating the sources and assumptions embedded in our models and algorithms, we risk perpetuating inequalities and overlooking vital perspectives.

If no district is to be left behind, then every district must be empowered to speak, to be heard and to act.

Nashin Mahtani is the director of Yayasan Peta Bencana [Disaster Map Foundation] in Indonesia, the MapaKalamidad Foundation in the Philippines, and the founding director of the international foundation, Climate Emergency Software Alliance (CESA). She leads a multi-disciplinary team in the development of one of the largest open source software projects for community-led climate adaptation, currently serving over 400 million people at the frontline of the climate crisis. In 2022, Nashin was the Indonesian youth delegate to the G20, advocating for climate and social justice at policy levels with global leaders. Nashin's contributions have earned her places on lists such as Forbes Asia 30 Under 30 in 2021 and nominations for the Global Citizen Prize: Cisco Youth Leadership Award.