Paradigm shifts don’t come around often, much less at such a rapid pace. But in the year between the International Association of Fire Chiefs’ (IAFC) first Technology Summit International in October 2022 and its subsequent event in December 2023, a game-changing technology erupted into public consciousness and is already transforming the way we live and work.
ChatGPT (GPT being an apt acronym for “generative pretrained transformer”), OpenAI’s artificial intelligence (AI) chatbot, gained 1 million users within five days of launching in November 2022. On ChatGPT’s first anniversary, it had more than 1.7 billion users, making it the fastest-growing consumer application in history. Although it began as a model for learning how people use and interact with generative AI, ChatGPT has quickly evolved into a deep-learning model that can not only recognize speech and images based on the data on which it was trained, but it can generate new content based on what it’s learned.
While ChatGPT is just the most visible recent development, AI has existed since the early 1950s, when researchers began trying to teach computers to mimic the problem-solving skills of a human. In 1996, the IBM supercomputer Deep Blue defeated world champion Garry Kasparov in the first game of a highly publicized chess tournament, representing the first public display of a computer’s decision-making abilities.
Preet Bassi, CEO of the Center for Public Safety Excellence (CPSE), who moderated a session on AI at the Tech Summit, explained AI in its simplest form: “Artificial intelligence is a field which combines computer science and robust data sets to enable problem solving. Even simpler, AI is a tool that automates either generation or prediction. AI systems can take two approaches: a human approach, where they think and act just like humans, or an ideal approach, where they actually think and act rationally.”
Given the hubbub around ChatGPT, AI was a major theme at this year’s Tech Summit, which gathered an impressive roster of tech-forward leaders in the fire service. Following are just a few ways these leaders acknowledged the use of AI in the fire service as well as the increased focus on better data and analytics tracking, and how other new technology is changing the industry.
Fire risk assessment
One area where AI has already impacted the fire service is risk assessment. AI can analyze data from disparate sources to identify areas of higher risk of fire, which then informs mitigation efforts.
Data from public and private sources range from building and occupancy permits, hazardous-materials storage, property records that indicate the building’s age, layout and design, construction materials, proximity to water sources and inspection reports to geographical and historical climate data. AI can analyze the data and generate risk assessment scores that departments can use to prioritize mitigation efforts from deploying firewatch drones, camera systems and lightning detectors in wildfire-prone areas to conducing fire-safety public information campaigns.
Preplanning
Fire service leaders addressed how strong data science and analysis can help with pre-incident planning so when a critical incident happens, fire-rescue crews have critical information at hand. From building plans and gate codes to the locations of electrical boxes, hazardous materials and hydrants, to mapping alternative routes, data-driven preplans help departments prepare for potential emergency situations so crews can spring into action without delay when the call comes.
Incident management
Despite all the preparation and preplanning a department can do, the nature of fire is unpredictable. Especially in a widespread incident like a wildfire or flood, multiple agencies may be called to respond. Gaining a common operating picture (COP) of the incident is critical to executing an efficient and effective response.
A COP uses a range of data sources to inform tactics and decisions:
- Incident reports that give critical early information about the size and scope of the incident.
- Location of available resources.
- Real-time weather data, including wind speed and direction, temperature, humidity and precipitation.
- Geographical information systems (GIS) data that provides details about topographical information like infrastructure, land use and population density.
- Satellite and aerial imagery giving a broad overview of the incident. Some aerial sources can provide real-time data on smoke composition and air quality.
- Fire modeling data that can predict fire behavior-based inputs like ignition source, smoke production and spread.
Firefighter safety
In addition to numerous applications for fireground situational awareness, advanced technology is increasingly being incorporated into firefighting gear. Some of the newest smart SCBAs have integrated sensors that can monitor the supply of breathable air and send alerts when the air supply is running low or if they detect specific toxins or gas. Some PASS devices also have air-quality sensors and will send alerts if the firefighter becomes immobile.
Smart helmets – a popular demonstration at the Tech Summit – are using thermal cameras, toxicity sensors and edge detection in an augmented reality optics to help firefighters see through dark smoke and navigate in low-visibility environments.
Community risk reduction
Michigan had one of the highest fire fatality rates in the nation. By thorough analysis of fire incident data reporting, State Fire Marshal Kevin Sehlmeyer gained insights that led to a statewide CRR plan that is moving the needle in the right direction.
In the session “Bridging the Data Gap: Turning Data into Actionable Intelligence,” Sehlmeyer explained that when the team started looking at fire fatality data to uncover trends, they learned that smoking was the top cause of fatal fires, 25% of fatalities were persons with disabilities, and 10% were on home oxygen. Most fires started in the kitchen or living room, and many victims died within five or 10 feet of an exit. Most fatalities didn’t have a working smoke alarm, “working” being a key phrase that has convinced Sehlmeyer that fire alarms need 10-year batteries.
Sehlmeyer used fire fatality data to effect change, including developing a web-based mobile app for fire departments to submit fatal fire data and document smoke alarm installation, and launching CRR efforts to address knowledge gaps. Soon, the app will be moving to the Esri platform, which will enable geographical mapping of fatal fire incidents statewide, in real time.
“We’re taking some would say very basic data on fatal fires, and we’re seeing trends, and we’re going to be able to make differences,” Sehlmeyer said. “One of the things we’re looking to do once we get Esri up and running is that we’re going to be able to then take all the alarms we’ve installed – those 17,000 homes we’ve been in – and then come back after the fact, take the CAD data from all of those and automatically determine how many of those homes we put up alarms in where we actually had a positive effect.”
Bart Van Leeuwen, a Dutch firefighter and technologist, also sees the value of gathering incident data in helping departments identify trends and document their successes.
“We collect incident data, but if you want to create stories for the public to understand, creating narratives out of your incidents will actually help to talk about in community outreach,” Van Leeuwen shared in the session “Safely Harnessing the Potential of Artificial Intelligence for Public Safety.”
Put simply, “We have to tell a better story and the data will lead us to that better story,” Sehlmeyer said.
Responsible use of AI
Many of the speakers at the Tech Summit acknowledged that AI is still a scary proposition for many, but consumer-friendly applications, like asking ChatGPT to explain a concept to a kindergartener or make recommendations for a surf trip to Costa Rica, are yielding helpful information for us to consider, but that we – especially those in public safety or other fields where decisions have life-or-death consequences – should never abdicate decision-making or responsibility to AI. After all, everyone in the fire service knows fire is unpredictable. AI can’t predict per se, but it can measure risk so we have more real-time information on which to base decisions.
“The biggest elephant in the room is the ethical and legal implications of using AI on the emergency response team,” Van Leeuwen said. “Who is responsible for the results? What if I make a decision based on an AI prompt and two people die? I have to live with the fact that people died based on the decision from AI.”
In fact, Van Leeuwen conducted research that found AI “can actually undermine decision-making skills it is supposed to improve, including forms of complacency, deskilling and avoidance of responsibility.”
Another big fear with AI – not just in public safety – is that bad actors are no longer hacking databases but instead using their nefarious talents to feed inaccurate information for AI systems to train on, generating misinformation. This has tremendous ramifications across the geopolitical world.
Fortunately, one of the major evolutions in ChatGPT and similar models is the marriage of OpenAI with cloud security. Sai Narain, director of technology, state and local government for Microsoft, explained: “LLM is a large language model that’s been trained on a large language dataset. GPT is one of those models built by OpenAI … Microsoft has taken that and built out an open AI service that sits inside its cloud so that you can now use that same technology that OpenAI has built in a manner that conforms to your enterprise security rules, data classification, data protection, etc., so you’re not leaking information out to the internet.”
Acknowledging the growth of AI and potential risks – both to caution users about technologies that can potentially wreak havoc and to allay fears about technologies that are already bringing tremendous benefits to solving global challenges – on Oct. 30, President Joe Biden issued a landmark executive order providing guidance for the responsible use of AI. The executive order “establishes new standards for AI safety and security, protects Americans’ privacy, advances equity and civil rights, stands up for consumers and workers, promotes innovation and competition, advances American leadership around the world, and more.”
The future of data in the fire service: NERIS
Under the leadership of Fire Administrator Dr. Lori Moore-Merrell, the U.S. Fire Administration (USFA) is leading a national effort to develop and launch a new interoperable fire information and innovative analytics platform, known as the National Emergency Response Information System (NERIS), which will ingest data from multiple sources via an application program interface (API) that will allow multiple software applications to talk to each other.
NERIS will eventually transform the National Fire Incident Reporting System (NFIRS) into an AI-powered secure, cloud-hosted system that, according to the USFA, will “empower the fire and emergency services community by equipping them with an empirical basis for decision-making. It will provide the community with reliable predictive analytics to support enhanced preparedness and response to all-hazard incidents, wildland urban interface events, community risk reduction efforts, climate change threats and associated resilience and mitigation efforts, and future pandemic emergency response resource preparedness.”
NERIS will ingest data from the best available authoritative services and provide data out in multiple interoperable formats. This will enable the fire service to mine empirical, data-driven insights that can be used to improve disaster response and enhance local CRR efforts.
“If we are going to extract data from our current systems, we have to hunt for it,” Dr. Moore-Merrell explained. “We learn incident by incident … we can make good actionable decisions based on data collected from NERIS.”
Understandably, an industry steeped in tradition like the fire service may have some hesitation in relying on AI. But as several of the Tech Summit speakers argued, AI can also improve how first responders prepare for and respond to critical incidents. It won’t replace the experience and wisdom of those battling the fires and leading innovation, but it will redefine the future of the fire service in the face of unprecedented threats and challenges in both the built and natural environments.
“I encourage everyone to lean forward,” Moore-Merrell added. “This will keep us from falling into the status quo.”
