Issue 017: Seattle's Hidden 911 AI, 24/72 Schedule Study, Data Center Fire Risk, And The PTSD Number No One Wants To See
Issue 017 - June 23, 2026
Read time: 8 minutes
This week: the Seattle Times reveals an AI system listening to every 911 call in one of the largest US dispatch centers since 2023, with no council approval and no outcome data published. Plus the first hard data on the 24/72 schedule debate from a peer-reviewed study, Oklahoma State engineers building nanofabrics that block carcinogens at the fiber level, a tri-university team mapping fire risk inside AI hyperscale data centers, Penn State and Toyota redesigning lithium-ion electrolyte chemistry to stop thermal runaway before it starts, and a new meta-analysis explaining why firefighter PTSD goes untreated even when programs exist.
AI in Dispatch
A Newspaper Just Caught One Fire Department Using AI On Its 911 Calls. Dozens Of Others Are Doing It Too.
A Seattle Times investigation published June 17 revealed that the Seattle Fire Department has been running AI software from Danish company Corti on every incoming 911 call since December 2023. The system listens to live audio, identifies low-acuity calls, and sends dispatchers a real-time popup recommending diversion to a nurse line instead of a fire or EMS response. The department pays $260,000 per year. The city council was never asked. No outcome data has been published showing whether diverted callers received appropriate care.
This is not a Seattle problem. Corti and competitors are actively selling these tools into US 911 centers right now. Most cities have no policy requiring council review, no requirement to track health outcomes for diverted callers, and no public disclosure framework. The Seattle Times piece is the first detailed public-records-driven look at what happens when a fire department buys AI for the dispatch floor without a governance plan, and it is the template every reporter, city council member, and plaintiff's attorney will use to look at every other department running similar tools.
The take: If your department is piloting AI triage in 911 - or your CAD vendor just pitched it last quarter - you have a narrow window to get ahead of this story before it becomes your story. Three things to do this week: get the contract and the system specs in front of your city attorney, ask your medical director for the outcome-tracking plan in writing, and brief your council on the program before a local reporter does. The technology is not the problem. The accountability vacuum around it is.
Read the full investigation at The Seattle Times →
Workforce
First Peer-Reviewed Data On 24/72 Schedules: Every Administrator Surveyed Was Satisfied. The Catch Is What You Already Suspected.
A study published June 4 in Frontiers in Environmental Health surveyed fire department administrators who had already converted to the 24/72 schedule, also called D shift, where firefighters work 24 hours on and 72 hours off across a four-platoon rotation. One hundred percent of respondents reported being satisfied or very satisfied with the change. Administrators cited improved morale, stronger recruitment, better retention, reduced sick leave, and improved shift coverage as the operational outcomes.
The catch is what makes most chiefs hesitate. The 24/72 model requires a fourth platoon, which is roughly a 33 percent increase in personnel headcount to maintain the same daily staffing levels. The most frequently reported challenge was increased overtime costs during the transition window, before the new platoon could be fully hired and trained. Departments that planned for the hiring lift in advance reported a smoother conversion than those that flipped the schedule first and tried to backfill later.
The take: Every union in the country is pushing 24/72 and a lot of chiefs have been pushing back on cost grounds with no peer-reviewed data to point at. This study is the first administrator-focused evidence in the literature that the schedule works organizationally, but it also quantifies the part nobody likes to say out loud - you need a fourth platoon, you need to fund the hiring before you flip the schedule, and you need to plan for transition overtime. Bring this to your next contract negotiation with eyes open on both sides.
Read the full study at Frontiers →
PPE Science
Oklahoma State Is Building Wildland Turnouts That Block Carcinogens At The Fiber Level
Researchers at the Textile and Apparel Science Laboratory at Oklahoma State University, writing in The Conversation on June 9, laid out a multi-year program to engineer nanofiber wildland gear that physically blocks ultrafine smoke particles from reaching skin while maintaining airflow for heat regulation. The trigger for the work is recent occupational health data showing firefighters face a 58 percent higher risk of dying from skin cancer than non-firefighters and a 40 percent higher risk of dying from kidney cancer. The OSU team visited active controlled-burn sites, collected dirty gear from real crews, and confirmed that fine carcinogenic particles were passing through many of the breathable fabrics currently in wildland use.
The OSU approach uses multilayer textile systems built from electrospun nanofibers, nanocoatings, and nanocomposite materials. The structures physically capture particles without trapping body heat. The team is deliberately avoiding PFAS chemistry because PFAS carries its own cancer risk and faces bans in turnout gear starting in 2027. One open question the researchers flagged: rougher fabric surfaces trap particles better but require improved cleaning procedures to keep contamination out of the skin contact layer over time.
The take: This is the long game on wildland cancer exposure. We are years away from a deployable garment, but the work is happening now and it is being done by a textile engineering team rather than a fire-gear vendor. That matters because the next generation of wildland standards will likely write the performance requirements before the products exist. Departments equipping wildland crews should be tracking this research and building their own gear-decon SOGs in parallel - the skin contact layer is where the next standards fight is going.
Read the full story at Phys.org →
Data Center Fire Risk
Three Universities Just Mapped Fire Risk Inside AI Data Centers. Departments In Data Center Corridors Should Read It Now.
A collaborative study from Texas A&M, George Washington University, and UC Berkeley, published in the Journal of Loss Prevention in the Process Industries and covered by TechXplore on May 26, analyzes the fire risk landscape inside modern AI hyperscale data centers. The team, led by Texas A&M chemical engineering professor Qingsheng Wang, found that data center fires typically originate from battery storage failures, electrical arc flashes, equipment malfunctions, and human error - and the dangerous finding is that these causes overlap and cascade. Human error during battery installation can trigger arc flashes that ignite other components. The same arc-flash outcome can also occur without human error from a simple short circuit.
The researchers concluded that suppression systems designed for legacy data centers may not be appropriate for high-density battery storage and the cooling configurations used in AI hyperscale facilities. They called for holistic fire prevention strategies, better fire incident reporting from the industry, and specialized suppression that does not destroy the equipment it is trying to protect. The context for fire departments is immediate: in Jerome Township Ohio, two Amazon data centers generated 84 emergency calls in four years, including a two-alarm fire in April 2026 that caused over fifty million dollars in damage and tied up local crews for more than 24 hours.
The take: Data centers are being approved at scale in jurisdictions that have not seriously thought about how to fight the fires they create. Virginia, Ohio, Texas, Arizona, and Iowa are stacking facilities right now. Departments in those corridors should pull this paper, walk it to the building official, and start the pre-incident planning conversation with the operator before the next 2 AM call. The structural fire toolkit does not cover the failure modes the researchers describe.
Read the full story at TechXplore →
Battery Safety Research
Penn State And Toyota Are Trying To Stop Lithium Battery Fires Before They Start - By Redesigning The Electrolyte Chemistry Itself
Penn State announced on June 16 that professor Chao-Yang Wang received a multi-year award from the Toyota Research Institute to develop a new generation of safer liquid electrolytes for lithium-ion batteries. The goal is to prevent thermal runaway at the chemistry level, before any suppression system has to engage. The work builds on findings Wang's team published in Nature Energy identifying the specific oxygen-scavenging mechanism that drives thermal runaway in failing cells.
The new program aims to synthesize electrolyte compounds that capture or block oxygen before it reaches the lithium anode, making the cell inherently less likely to catch fire. The applications extend beyond electric vehicles to AI data center battery storage systems and grid-scale storage installations, both of which fire departments are increasingly being called to. The award is part of Toyota Research Institute's largest university program cohort to date - 104 faculty across 31 universities. The work will run multi-year.
The take: The fire service is rightly focused downstream - new SOGs for EV fires, new suppression agents for battery storage, additional training hours on thermal runaway response. This is the upstream fix. If the chemistry works, the cells stop failing in the first place. That is years out from product, but the strategic point for chiefs is that watching only the suppression side of battery safety misses where the actual safety budget is moving. The next ten years of EV and BESS fire risk will be set by labs like this one, not by gear vendors.
Read the full story at Penn State News →
Deep Dig · Behavioral Health
A New Meta-Analysis Just Quantified Why Most Firefighter Mental Health Programs Fail To Reach The People Who Need Them
A peer-reviewed narrative review published May 30 in the Journal of Men's Health synthesized six years of research on PTSD and common mental disorders in male first responders, including firefighters and paramedics. The headline numbers are bracing on their own - PTSD prevalence ranges from 7 to 37 percent across studies, depression 15 to 25 percent, anxiety 15 to 20 percent, and hazardous alcohol use 25 to 30 percent. The 7 to 37 percent range is not uncertainty, it is a signal that departments without systematic screening are likely living near the high end of that range without knowing it.
The harder finding is the mechanism. The review found that the primary barrier to treatment is not access to programs. Most departments now have peer support and EAPs. The barrier is occupational culture and masculine role expectations that create stigma around help-seeking, which means the people who need care most often will not raise their hand to receive it. Evidence-based therapies including cognitive processing therapy, prolonged exposure, and EMDR showed large effect sizes for PTSD reduction when accessed. They are underutilized because the culture discourages self-identification. A separate 2026 Frontiers in Public Health network analysis added another layer - sleepiness is the central bridge variable that connects burnout, stress, and work-family conflict, which suggests shift-work sleep disruption is the mechanism through which the job damages mental health over time.
The take: The peer support booth at the trade show is not the problem and was never the problem. The problem is that the culture inside the firehouse penalizes the act of raising your hand. Chiefs who want to actually move the needle have to do two specific things this research backs: build systematic screening into the annual physical so it is not voluntary, and address sleep deprivation as a wellness intervention rather than a scheduling preference. The cancer registry shifted the culture on carcinogen exposure in ten years. Behavioral health is where that fight is now.
Read the full review in the Journal of Men's Health →
What We're Watching
City auditor reports on fire-based EMS systems are picking up speed. Several large US cities are running performance reviews of their EMS call volume and apparatus utilization. The next set of findings will likely surface in late summer and frame fall budget conversations. Expect more "transport-vs-respond" framing in city audits that will pressure dispatch and unit-deployment decisions.
Federal IG audits of fire department procurement. A June NYC Comptroller audit found FDNY paid over a million dollars per year for a fire inspection software platform that ran unsupported for two years. Expect similar audits in other major metros as enterprise software contracts come up for review, with particular focus on AI and cloud-based systems purchased without competitive bid.
Battery storage at scale is the next code fight. Battery Energy Storage Systems are being installed at residential, commercial, and utility scale faster than local code officials can write new requirements. NFPA 855 covers the basics but jurisdictions are starting to write supplemental local rules. Departments with rapid BESS growth should be at the building code adoption table now.
Sleep science as occupational health. Research linking shift-work sleep disruption to long-term mental and cardiovascular health outcomes is building in volume. The framing of sleep as an occupational hazard rather than a personal preference is gaining ground in the literature, which could shift how nap policies, station design, and shift schedules are evaluated in the next NFPA wellness update.
Subscribe to The Overhaul Brief
Fire service tech, AI, and innovation delivered every Tuesday.
