Skip to content
Furkan Bayıroğlu
Early-stage project · In development

AfetBridge

Connecting Needs, Resources and People After Disaster.

AfetBridge is an AI-assisted disaster response and coordination platform being developed to help communities and response teams organize needs, resources and field information after earthquakes and other disasters.

  • AI
  • Disaster Response
  • Civic Technology
  • Geospatial

This page presents the AfetBridge project and its intended design. It is not the AfetBridge application, and the system has not been deployed.

01The Problem

Critical information becomes fragmented.

After a disaster, information about needs, resources and conditions on the ground arrives through many separate channels at the same time.

Where information scatters

  • Messaging platforms
  • Phone calls
  • Social media
  • Spreadsheets
  • Volunteer groups
  • Field reports

What this can lead to

  1. 01Duplicate reports
  2. 02Outdated information
  3. 03Unstructured requests
  4. 04Difficulty identifying urgent needs
  5. 05Difficulty connecting resources with needs

02The Idea

Turning fragmented information into coordinated action.

AfetBridge is designed to help organize incoming information and make it easier for coordinators to understand what is happening.
  1. 01

    Report

    Information arrives as free-form messages and field updates from many different channels.

  2. 02

    Understand

    Reports are organized into consistent fields: what is needed, where, how urgently, and from which source.

  3. 03

    Verify

    Coordinators confirm, correct or reject information. Verification status stays visible at every step.

  4. 04

    Match

    Needs are compared with available resources, and possible matches are suggested for review.

  5. 05

    Coordinate

    Teams work from a shared, current picture instead of scattered messages and lists.

03AI Role

AI-assisted. Human-verified.

AI is intended to reduce the manual work of organizing information, so people can spend their attention on decisions.
  • Structuring free-form reports
  • Classification
  • Summarization
  • Multilingual communication
  • Identifying potential duplicate reports
  • Suggesting potential matches between needs and resources
  • Natural-language search across operational information

AI does not replace human decision-making in critical situations.

AI output is treated as a suggestion. It is labelled as unverified until a person reviews it, and its uncertainty is shown rather than hidden.

Incoming reportSource: message

“Need drinking water and blankets near the school on the main street. Around 20 people, some elderly.”

AI-structured draftUnverified · suggestion
type
need
items
drinking water, blankets
people
~20 (approximate)
location
near school, main street · needs confirmation
source
linked to original report
Coordinator reviewHuman decision
ConfirmCorrectReject
Illustrative example — not real data.

04Core Capabilities

What AfetBridge is designed to do.

  • Needs & Requests

    Collect and structure reports about urgent needs.

  • Resource Coordination

    Organize available supplies, volunteers and logistical resources.

  • Verification

    Distinguish clearly between unverified and verified information.

  • Operational Map

    Visualize needs, resources and important locations geographically.

  • Smart Matching

    Suggest possible relationships between needs and available resources.

  • Multilingual Support

    Help information move across language barriers.

  • AI Operations Assistant

    Allow coordinators to ask questions about operational information using natural language.

  • Capabilities describe the intended design and may change as development continues.

05Built for Difficult Conditions

Designed for difficult conditions.

Disasters disrupt networks, power and routines. These are the design goals that guide how AfetBridge is being built.

These are design goals for a system in development. AfetBridge has not been field-tested.

  1. 01

    Mobile-first

    Field information is mostly sent and read on phones, so the phone is the primary target.

  2. 02

    Low-bandwidth friendly

    Lightweight pages and small payloads for congested or degraded networks.

  3. 03

    Simple workflows

    Few steps to report or review, usable under stress and without prior training.

  4. 04

    Clear information hierarchy

    What is urgent, verified or out of date should be visible at a glance.

  5. 05

    Offline-friendly foundations

    An architecture that can be extended to hold reports while connectivity is unavailable.

  6. 06

    Resilient architecture

    Components separated so that a problem in one part does not take down the whole system.

  7. 07

    Transparent uncertainty

    Verification status, confidence and the age of information are shown rather than hidden.

06Responsible AI

Designed around responsible AI.

  • 01

    No fabricated certainty

    Unverified information remains clearly identified.

  • 02

    Human oversight

    Critical decisions remain under human control.

  • 03

    Traceable information

    AI interpretations should remain connected to their source reports.

  • 04

    Privacy by design

    Only necessary information should be collected and sensitive information should not be exposed publicly.

07Future Vision

Built for collaboration.

Disaster response is collective work. Over the long term, we want to explore working with the people and organizations who carry it out.
Discuss a collaborationAll projects

Potential future collaborators

  • Communities
  • Volunteers
  • NGOs
  • Municipalities
  • Humanitarian organizations
  • Technology partners

These are kinds of collaboration we hope to explore. They are not existing partnerships.

AfetBridge · Early-stage project · In development