top of page

SPONSOR- FIRST RESPONDER NETWORK AUTHORITY(FIRSTNET)

UX RESEARCH

TIMELINE- 4 MONTHS 

Designing AI Systems for Public Safety

01 PROJECT OVERVIEW

Firefighters make critical decisions in high-risk environments where visibility, communication, and situational awareness are often limited.This project explored how artificial intelligence could support emergency responders without disrupting trusted workflows or increasing cognitive load.

​

Rather than designing a single product, our team examined the broader AI ecosystem and developed evidence-based use cases that balanced innovation with the realities of emergency response.

The outcome was a collection of AI-powered opportunities and design recommendations intended to help public safety organizations understand where AI can provide value and how these systems should be designed for adoption.

1a. OUR APPROACH 

01. Secondary Research 

Explored existing tools, AI technologies, paint points and ethical considerations 

02. Interviews

Spoke with firefighter and fire chiefs to understand real world experience, needs and concerns  

03. Observations 

Visited 3 fire departments and observed training exercises and equipment in action

04. Synthesis 

Mapped insights and identified patterns, paint points and opportunity areas

05. Use cases

Developed AI use cases and design principles grounded in real-world workflows

1b. IMPACT AND RESULTS 

08

Use cases shipped 

03

Fire departments observed 

12+

Stakeholders/Firefighters interviews

40

Products screened 

02 PROBLEM DEFINING

Most conversations about AI in emergency response focus on technical capability.But firefighters do not adopt tools because they are technologically impressive.

They adopt tools that are:

  1. Reliable

  2. Fast to learn

  3. Easy to deploy

  4. Durable under pressure

  5. Compatible with existing workflows

 

The challenge was not simply:“What can AI do?”

 

It was: How can AI be designed to support firefighters in ways that feel dependable, practical, and operationally realistic?The project sought to understand both AI opportunity and human hesitation surrounding adoption.

PROBLEM STATEMENT

How might we design AI systems that improve firefighter safety and situational awareness while remaining intuitive, trustworthy, and compatible with existing emergency response workflows?

SECONDARY RESEARCH

To better understand tasks faced by firefighters on an active scene and how AI can be best used to support these efforts, desk research across various domains was conducted:

​

1. Existing non-AI powered tools for real-time firefighting usage

Tools researched included:

  1. Radios and communication devices

  2. GPS navigation systems

  3. Weather monitoring solutions

  4. Thermal imaging cameras

  5.  Personal protective equipment (PPE)

​

Takeaway- This aided in better understanding firefighter’s current toolset. It was used to guide future analysis as to how these products could be tailored to incorporate AI. â€‹

2. Privacy and ethical implications of AI in public safety/firefighting   

 

 

Articles examined offered a comprehensive understanding of the ethical, privacy, security, and operational implications of integrating AI technologies into firefighting.

​

a. Ethical concerns mount as AI takes bigger decision-making role in more industries

b. AI Technologies, Privacy, and Security

c. Summit explores new tech, AI and the shift to data-driven models

​

​

Takeaway- The results of this desk research suggested that any use cases recommended should be developed using AI algorithms that prioritize fairness and transparency in decision-making processes during firefighting operations, ensuring unbiased outcomes and accountability. 

3. Existing research surrounding AI in real-time firefighting scenarios 

​

These use case ideas were used as a starting point to guide the team’s further research into what kinds of AI products may already be out on the market, or in beta testing, that provide value to firefighters on an active scene.

​

Articles and studies examined were as follows:

a. Applying artificial intelligence (AI) to improve fire response activities

b.Li, J., Brown, C., Dzikowicz, D.J., Carey, M.G., Tam, W.C., & Huang, M.X. (2023).

​

Takeaway-  This aided in filling some knowledge gaps the research team had, guiding future research phases based on studies analyzed, and therefore helped work towards achieving the overall objective.

4. Existing AI products for real-time firefighting usage   

Explored AI applications involving:

  • Predictive analytics

  • Hazard detection

  • Drone mapping

  • Health monitoring

  • AR-assisted navigation

  • Fire behavior prediction

  • Real-time decision support

​

Takeaway- This aided in the goal of writing use cases for each product.

INTERVIEWS & OSERVATIONS

OSERVATIONS

To understand firefighters day-to-day realities, our team conducted observational visits to three fire departments in Greenwood, Indiana. These field visits helped uncover challenges and opportunities that were difficult to identify through interviews or desk research alone.

Research Activities

  1. Observed survival and rapid intervention training

  2. Explored firefighting trucks and EMS equipment

  3. Conducted live demonstrations and discussions with firefighters

  4. Studied daily workflows, team coordination, and technology use

Key Observations

  1. Communication is critical in emergency situations, especially in low-visibility environments

  2. Firefighters rely heavily on verbal coordination and teamwork

  3. Technologies like thermal imaging cameras and pack tracking systems play an essential role in safety and navigation

  4. Emergency environments require quick decision-making under high physical and mental stress

Challenges Identified

  1. Signal interference inside large or complex structures disrupted communication

  2. Limited visibility made navigation and situational awareness difficult

  3. Heavy equipment and protective gear created physical strain during operations

  4. Existing systems sometimes lacked seamless coordination in high-pressure scenarios

Design Insights & AI Opportunities

These observations directly shaped our AI concepts and ensured they reflected real operational needs.

  1. AI-enhanced communication to reduce signal interference and improve team coordination

  2. Real-time firefighter safety monitoring through smarter tracking systems

  3. Context-aware support tools to improve awareness and decision-making during emergencies

Impact

Grounding the design process in field research ensured the final AI solutions were practical, safety-focused, and aligned with firefighters’ real-world workflows.

WhatsApp Image May 18 2026 (2).jpeg

INTERVIEWS

Moderator-led interviews with firefighters helped our team understand their experiences, operational challenges, and perspectives on AI in emergency response. These conversations explored technology use, ethical concerns, and opportunities for AI to support firefighters during active scenes.

The questions explored:

  1. Active scene challenges

  2. AI familiarity

  3. Safety implications

  4. Privacy concerns

  5. Desired UX characteristics

  6. Integration with current tools

What firefighters told us

"Technology needs to be reliable and simple to use when we’re under pressure.”

                -Firefighter

“We prefer improving tools we already know rather than learning completely new systems.”

"If AI is introduced, it should work with our training and existing equipment not replace everything we already use.”             

         -Fire chief

RESULTS/FINDINGS 

Key Challenges 

Ethical & Privacy considerations

Firefighter operate  chaotic, unpredictable and high-risk environments 

They face communication barriers, limited visibility and incomplete information

While AI holds promise, adoption is limited due to usability, reliability and ethical concerns 

1. Minimal data collection- Collect only whats essential 

​

2. Local processing where possible- Reduce reliance on constant connectivity 

​

3. Transparency & explainability - show sources, confidence and reasoning

​

4. String security & encryption- Protect crews and mission-critical data  

KEY INSIGHTS

RELIABILITY OVER INNOVATION-

Firefighters value tools that are dependable, easy to set up, and stable under pressure

Visibility, navigation, hazard detection, and tracking are the biggest pain points 

SITUATIONAL AWARENESS IS CRITICAL-

ADOPTION DEPENDS ON FAMILIARITY-

Tools must integrate into existing workflows and reduce-not add-complexity 

ETHICAL TRUST IS NON-NEGOTIABLE -

Privacy, data ownership, and transparency are major concerns around AI adoption

SOLUTION EXPLORATION

After synthesizing research insights, we shifted focus from identifying problems to exploring where AI could realistically support firefighter workflows without introducing friction or risk.Rather than designing a single “all-in-one” product, we evaluated multiple emerging AI systems and mapped them to specific stages of emergency response.

Design Direction: AI as Augmentation, Not Replacement

 

A key principle emerged from research:

Firefighters do not need more systems to learn. They need better intelligence inside systems they already trust.

 

This led to a design direction focused on:

  1. Embedding intelligence into familiar workflows

  2. Reducing cognitive load during high-stress moments

  3. Improving real-time awareness without adding complexity

  4. Supporting decision-making, not replacing it

 

AI was positioned as a silent support layer, not a dominant interface.

USE CASES 

 1

LIDAR DRONE MAPPING 

Real-time 3D mapping of structures and terrain to identify hotspots, collapse risks and safe entry routes 

Value to firefighters:

  1. Rapid building structure visualization

  2. Identification of hotspots and collapse risks

  3. Safer perimeter assessment

  4. Improved incident command awareness

 

Design insight:Drones are most effective when they act as remote sensory extensions, feeding simplified insights—not raw data streams—to incident commanders.

Aerial situational awareness

 2

AUDREY(AI DECISION SUPPORT)(Assistant for Understanding Data through Reasoning, Extraction and Synthesis)

Uses Machine learning tools to improve data from sensors, thermal cameras and tracking system to surface actionable insights and risk indicators 

Value to firefighters:

  1. Aggregated situational awareness

  2. Early hazard detection

  3. Pattern recognition in fire spread

  4. Resource allocation support

 

Design insight:The success of systems like AUDREY depends on confidence transparency—firefighters need to understand why the system is making a suggestion, not just what it suggests.

Decision support 

Derived from https://www.govexec.com/media/nvidia-gai-delltechnologies-aicasestudy.pdf

 3

C-THRU by Qwake Technologies

Value to firefighters:

  1. Obstacle detection in low visibility

  2. Navigation assistance inside structures

  3. Highlighting exit routes and hazards

  4. Reducing disorientation risk

​

Design insight: AR systems must be extremely minimal—too much visual data can increase cognitive overload during emergencies.

AR helmet mounted wearable device that highlight obstacles, exits and hazard in low-visibility, smoke filed environments 

Navigation Assistance  

 4

FIRESCOUT AI (AUTONOMOUS MONITORING)

https://firescout.ai/how-it-works/

Continuously monitors fire behavior and environmental changes to predict spread and escalators 

Value to firefighters:

  1. Continuous fire tracking

  2. Predictive spread modeling

  3. Real-time environmental updates

  4. Early warning for escalation

​

Design insight: Predictive tools must be treated as probabilistic guidance, not deterministic truth, to avoid over reliance.

Predictive Monitoring

 5

FIRE NEURAL NETWORK (WILDFIRE AI SYSTEM)

https://firescout.ai/how-it-works/

Value to firefighters:

  1. Fire behavior prediction

  2. Risk zone identification

  3. Resource planning support

  4. Evacuation strategy insights

​

Design insight: Model outputs must be simplified into actionable risk levels, not complex analytics dashboards.

Machine learning models predict fire behavior using weather, terrain and historical wildfire data

Risk Prediction

bottom of page