Industries / Government, education and community / Government
Emergency services
Emergency services answer calls for help and send crews to put out fires, treat the sick and injured where they are found, and rescue people who are trapped, lost or in danger.
6 CE
After fires destroyed many parts of Rome, the emperor Augustus formed a company of freedmen in seven divisions to fight them, meaning to disband it soon; he kept it, and the historian Cassius Dio wrote two centuries later that these night watchmen, the vigiles, still had barracks in the city and drew pay from the public treasury.
46
Milestones achieved across 7 domains.
11
To watch for: announced, not yet confirmed outside.
15
Open deliverables. Nobody has shipped them.
In one minute
Where it started
Rome paid a standing corps to fight fires from 6 CE. Fire brigades run and paid for by towns spread from Edinburgh in 1824 and Cincinnati in 1853, a charity for lifeboats was founded in London the same year as Edinburgh's brigade, and in 1967 Pittsburgh trained some of the first crews who treated patients at the scene, later called paramedics. Many of the rules crews work by today were written after deaths, from Mann Gulch in 1949 to Charleston in 2007.
Where it stands
In most countries a single number reaches the emergency center, 112 across Europe since 1991, and dispatchers coach callers through chest compressions and send trained volunteers nearby by phone. Crews carry breathing sets and thermal cameras, use tactics tested on burning houses, and get distress calls relayed by satellite from phones and beacons. The weak points are coordination and warning: radios that failed in New York in 2001, a "stay put" order held too long at Grenfell Tower in 2017, alerts that did not reach residents in Los Angeles County in 2025, and the crews' own health, now that firefighting is classed as a cause of cancer.
What is open
Most open jobs here are public tests and datasets built from published records, staged calls, simulations and training grounds, never a live call or incident. Among them: scoring AI interpreters and note-taking tools on staged calls, coding NIOSH's reports on more than 700 US firefighter deaths into one table, measuring how well drones spot people on a test course, and checking evacuation models against what NIST recorded in Paradise in 2018. AI models that transcribe speech, read documents and recognize images put these jobs within reach of a small team.
Domains
AchievedWatch forOpen
Domain
Progress
Last achieved
Watch for
Open
Taking calls and dispatch
Answering emergency calls, working out what is wrong and where, sending the right crews, and keeping in touch with them by radio and data link.
7 · 2 · 2
Last achievedCars call 112 themselves2018
Watch forAI interpreters on 911 calls
2
openFighting building fires
Putting out fires in homes, shops, factories and vehicles once the crews arrive, and getting the people inside out alive.
7 · 1 · 2
Last achievedBatteries that reignite2020
Watch forRescue sheets that list hazards
2
openWildfire response
Fighting wildland fires, the trade's term for fires in forests, grass and bush, and moving people out of their path.
7 · 2 · 2
Last achievedAlerts that did not arrive2025
Watch forSmoke cameras before 911 calls
2
openParamedic care
The treatment paramedics and other ambulance clinicians give at the scene, and the trials and rules that decide which treatments they carry.
6 · 1 · 2
Last achievedBone or vein, same result2024
Watch forAI notes on ambulance tablets
2
openSearch and rescue
Finding people who are lost, trapped or in distress, at sea, in mountains and wild country, underground and in collapsed buildings, and getting them out.
7 · 3 · 3
Last achievedSatellite SOS from a phone2022
Watch forSecond-generation distress beacons
3
openDisaster response
Organizing many agencies, countries and volunteers when an earthquake, storm, flood or fire is too big for local services, from the first hours to counting the dead.
6 · 1 · 2
Last achievedCounting the dead again2018
Watch forJapan's new disaster agency
2
openTraining and equipment
Teaching responders the job safely, and the gear that keeps them alive while they do it: breathing sets, protective clothing and cameras.
6 · 1 · 2
Last achievedChemicals measured in the gear2024
Watch forGear made without added PFAS
2
open
Narrower industries
Open in emergency services
3 of 15 open deliverables
Open · Taking calls and dispatch
Emergency calls through AI interpreters
A public set of scripted emergency calls in at least 10 languages, spoken by native speakers and by synthetic voices, with background noise and panicked delivery. Each script carries a fixed list of key facts: address, callback number, what happened, how many people are hurt, and whether the patient is awake and breathing. Run every call through several AI interpreters and score how many key facts reach the dispatcher's language correctly and how fast. Use only test systems or recordings, never a live emergency line.
Built with what exists
speech-to-text models such as the open-weight Whisper or ElevenLabs Scribe v3 (see the Note "Fewer word errors"); open translation models and general AI models; text-to-speech voices; public noise recordings; the NineOneOne paper's labels for call content as a starting list of key facts.
Open · Fighting building fires
Score every EV response guide
A public scorecard that grades the emergency response guide for every electric and plug-in hybrid model on sale in the United States against the NTSB's published yes/no checklist: ISO 17840 layout, steps to disconnect the high-voltage battery, how to stabilize the car, general battery-fire warnings, vehicle-specific instructions on where and how to apply water, submerged vehicles, towing and storage. Re-run it every year so readers can see which makers improve.
Built with what exists
makers' published guides; pdftotext; a language model that answers fixed yes/no questions about each document, such as Gemini Decision 1 (see the Note "A thousand questions, one pass"); two people checking a sample of the answers.
Open · Wildfire response
Score evacuation models on Paradise
A public test that runs several open traffic evacuation models on the roads of Paradise, Concow and Magalia, using the evacuation orders, road closures and burnovers NIST recorded for the Camp Fire, and scores each model's traffic hour by hour against what NIST observed. A burnover is a fire passing over people or vehicles.
Built with what exists
Eclipse SUMO, an open-source traffic simulator; WUI-NITY, a wildfire evacuation platform described by Wahlqvist, Ronchi and colleagues in 2021; OpenStreetMap road data; NIST's observation tables, read into a spreadsheet.
Where it sits
Broader
In outline