AI Report Generation for First Responders
Our research focuses on the exploration of AI technology, firefighting processes, and critical information capture necessary during wildfire incidents.
Role
UX Researcher
Timeline
2023 - 10 Weeks
​
Project type
Exploratory Project
Team
3 Researchers
Overview
"How might we use AI technology to assist in report generation?"
Our research focuses on the exploration of AI technology, firefighting processes, and critical information capture necessary during wildfire incidents.
The original brief for this project focused on the exploration of AI Technology with voice input to aid EMS with report generation. However, after meeting with the client we realized that the project was exploratory.
Approach
We segmented the project into four chunks.
Focus vs Exploration
Narrowing focus and targeting areas for exploratory research.
Methods
Understanding & validating the problem
Desk Research
Literature Review
We split Desk Research into 5 areas
Artificial Intelligence how does it work?
Past Firefighting Technology
Past Technology created for firefighters was:
-
costly
-
complicated
-
time-consuming to learn and use
-
not enough accuracy
-
extra equipment can get in the way
Focus on current equipment and technology and how we might repurpose it as opposed to creating a new technology
What do firefighters have on their uniform that we could repurpose?
Current fire incident report processes and reports
Assumption 1
Reporting is primarily through paper or low-tech reporting:
-
Firefighters still use paper for activities like incident command: paper maps for public buildings, paper manuals
​
Assumption 2
We should find digital "papers" reports
online:
-
We found 1 paper template of the report​
-
Valuable information regarding the reporting process was captured during the Head Quarters and Station Tour
WildFire Incident
A series of 16 tragic wildfires, occurring between 1937 and 1956 prompted the USDA Forest Service to create 10 Standard Firefighting orders, developed in 1957. These 10 Standard Firefighting Orders were developed to manage the unpredictable nature of wildfires and save the lives of personnel working under these conditions.
The Anatomy of a Wildfire
Source: South King County Fire Training Consortium: Wildland Urban Interface Awareness
Anchor Points
Insights
Situational awareness is key to managing wildfire incidents.
-
We need to know what the fire is doing at all times.
-
How does wind affect fire movement?
-
Where are safety zones?
-
Who is on-site and where are they?
-
What resources are available?​
New Questions emerge:
-
What kinds of input data could we collect on situational awareness and environmental conditions that would benefit firefighters?
-
Could we capture satellite images of the wildfire and label the main components of the fire? (head, heel, fingers, etc.)
-
How might we capture and display constantly changing weather conditions?
Subject Matter Expert Interviews: Insights
Observation / Contextual Inquiry
What we learned.....
-
Fire administration has strong reporting systems in place
-
While missing reports are an issue, there is a strong management system in place and automated notifications are sent to those who are missing reports
-
Reports are not time sensitive and Incident Commanders have up to 3 days to submit reports
-
Reporting done after the incident not during
-
Manual reporting did not seem to affect critical aspects of the ongoing incident
-
Reporting narratives are used as evidence in court cases and for insurance purposes
Data gathered from this method pushed us in a new direction:
Away from AI Report generation and towards a focus on non-traditional reporting of environmental factors like; wind-tracking, weather tracking, fire tracking, and how we might capture those elements.
Affinity Diagram
Insights from Affinity Diagram
Results
Initial Goals Versus Results
Next Steps
DESK RESEARCH: PHASE 2
-
Suggest researching AI and wind tracking, AI and location tracking, and AI and heat tracking.
-
Explore inputs for an AI algorithm that could track fire movement, location, and how wind might affect the direction of the fire.
-
testing and validation with firefighters
-
Further exploration of heads-up displays for air-pack masks, smart glasses, and smart goggles format.
-
Initial prototypes and sketches
