Executive summary
The landscape of physical safety has transformed dramatically over the past decade. Security once meant locks, alarms and human vigilance, but today’s safety ecosystem encompasses AI-powered cameras that spot weapons in crowds, drones that find missing persons using thermal imaging, and smartwatches that automatically call for help when we can’t.
- Police drone usage has exploded from 3% to 31% between 2015-2024.
- Smart home security adoption reached 73% of US households in 2024, with 50% using cameras.
- LASD used a heat-seeking drone to locate a missing man with dementia.
- A smartwatch saved a man’s life by automatically dialling 911 after he fell and remained motionless on the ground.
- The global public safety software market will double to $36.47 billion by 2030.
- Privacy concerns grow as 85 million surveillance cameras monitor US streets.
These advances arrive at a critical moment. With urban crime concerns persisting and emergency services stretched thin, technology offers powerful new capabilities. Yet as surveillance becomes ubiquitous and AI makes life-or-death decisions, society faces profound questions about privacy, bias and the balance between safety and freedom.
The world we live in isn’t exactly what you’d call a safe place. Turn on the news or scroll through social media, and you’ll see countless stories that will make you think twice about walking alone at night or leaving your car unlocked. What makes this reality even more unsettling is that it’s not getting better – it’s actually getting worse. Criminals today are more sophisticated than ever, constantly adapting their methods and finding new ways to exploit our vulnerabilities. Every technological advance that makes our lives easier seems to hand criminals a new tool for exploitation. Yet for every vulnerability it creates, technology also offers new forms of protection.
The good news is that AI and connected sensors are quietly changing how we protect human life, acting like digital guardians that work tirelessly in the background to keep us safe. Acoustic sensors can detect gunshots and alert police within seconds. Thermal imaging on drones can find missing persons in minutes rather than hours. Smart watches can detect when someone has fallen and automatically summon help. Across the globe, these technologies are already saving thousands of lives by enabling faster response times and early intervention.
The evolution of safety tech
From early CCTV networks to modern AI systems, safety technology has become an inextricable part of urban environments.
The journey towards modern safety systems has spanned the better part of a century. Early examples of safety tech include closed-circuit television (CCTV) networks, which went live in the mid-20th century, and the introduction of the 911 emergency call system, both of which laid the groundwork for rapid emergency response. These innovations shared a common thread: they extended human capabilities rather than replacing them. A security guard could now monitor multiple locations through CCTV, while dispatchers could coordinate emergency response across entire cities.
In policing, technologies like two-way radios, the CompStat crime-mapping system introduced in the 1990s, and widespread surveillance cameras have proven transformative. CompStat, in particular, showed how data aggregation could reveal patterns that weren’t immediately evident to individual officers. Suddenly, police departments weren’t just reacting to crime. They could see where problems clustered and when they peaked. It was the beginning of predictive policing, though nobody called it that yet.
Fast forward to today, and those old CCTV systems have evolved into something far more sophisticated. While the fundamental mission hasn’t changed much – spot danger, respond fast, maybe even prevent it – the speed, scale, and autonomy with which technology can now pursue that goal have transformed completely. In the past, a security guard might miss something crucial while checking another monitor. Now, AI systems process multiple video feeds simultaneously, catching details that human eyes would miss. They can spot a weapon in a crowded subway station and alert authorities before anyone else notices the threat.
The current state of safety
Cities worldwide are implementing vast networks of connected eyes and ears, capable of detecting threats in real time.
Of course, this transformation didn’t happen overnight: according to a University of Michigan survey, body-worn camera adoption by police jumped from 23% in 2015 to 77% in 2024. Similarly, police drone use increased from 3% to 31% over the same period. Unfortunately, true AI adoption remains in its infancy – only 3% of agencies currently use AI or predictive policing software, although one-third reported they were considering it.
Meanwhile, our infrastructure is getting smarter. An estimated 85 million surveillance cameras are now deployed across the US, monitoring streets, transit systems, and public venues around the clock. In the past, such devices were passive recording devices, but today they go a great deal further. In New York, for example, authorities used facial recognition and video analytics in late 2024 to track a murder suspect across thousands of networked cameras, leading to a quick arrest.
The technology’s impact on emergency response times is particularly striking. According to the National Emergency Number Association, reducing response time by just one minute could save 10,000 lives annually in the US. In recognition of this stark figure, emergency services are adopting AI and automation tools to compress those critical windows when survival hangs delicately in the balance. In the last couple of years, numerous cities across the US have implemented intelligent traffic signal systems that give priority to ambulances and fire trucks, cutting emergency vehicle travel times by up to a staggering 69%.
Naturally, the transformation also extends into emergency call centres, where AI is revolutionising how emergencies are processed and prioritised. In Jefferson County, Colorado, an AI ‘virtual dispatcher’ system reduced call processing times so significantly that a once chronically understaffed centre met call answer-time standards for six months in 2023, after 22 straight months of failing to do so. The AI transcribes calls instantly, identifies key information, and can even triage resources before human dispatchers take over.
The difference between life and death
From heat-seeking drones to gunshot detection systems, safety tech is demonstrating its effectiveness in real-world situations.
The statistics are compelling, but it is, of course, the individual stories that bring home the reality of what this technology offers. Take, for instance, the case of a 78-year-old man with dementia who wandered from his Malibu home on a cold December evening in 2024. With temperatures dropping to around 5°C, the Los Angeles County Sheriff’s Department (LASD) launched an urgent search. Unfortunately, ground teams struggled to find the missing man in the dark, wooded terrain. That’s when deputies deployed a heat-seeking drone equipped with thermal imaging.
Within 80 minutes of the family’s 911 call, the drone’s infrared camera had picked up a human heat signature lying in a brush-filled ravine about a quarter mile from the man’s house. He was scraped, shivering – a little worse for wear – but fortunately alive. “The operation highlighted the importance of swift action, teamwork, and innovative technology in locating missing persons, especially those vulnerable due to medical conditions,” LASD said in a statement. Captain Jim Braden of LASD noted that the drone “quite possibly saved the man’s life by finding him so quickly.”
This is far from an isolated incident. In 2024 alone, LASD drones were deployed in 12 search-and-rescue operations and 45 other high-risk incidents (like armed standoffs or disasters). Similar success stories have been reported nationwide. Maryland officers were able to find a missing autistic child in the woods in 15 minutes using thermal imaging. In North Carolina, drones located elderly hurricane victims trapped by floodwaters in the wake of the devastating 2024 hurricane. In Nevada, rescuers sent water and a radio to stranded hikers via drone, while Utah teams used drones to drop flotation devices to a child caught in river currents.
Chicago’s gunshot detection system shows AI’s life-saving potential at scale. From January 2023 to July 2024, ShotSpotter alerts enabled first responders to reach and render aid to 1,935 shooting victims who might otherwise have bled out. The system works by triangulating gunshots within seconds and sending alerts to officers’ smartphones, often before any 911 call is made. Analysis has found a 19% reduction in firearm homicides in districts after implementing the technology. One case in particular illustrates the palpable difference SpotShotter can make: in April 2024, sensors picked up a 15-round volley of gunfire in the city’s Austin neighbourhood at 2:30 AM. Officers were on scene in less than three minutes, finding a 27-year-old victim bleeding from multiple wounds. He survived – an outcome that might not have occurred without the rapid response enabled by the technology.
Over in Australia, the Lynxight AI drowning-detection system made headlines in June 2025 when it enabled Western Australia’s first AI-assisted pool rescue. The system alerted a lifeguard 21 seconds after a swimmer slipped underwater due to a medical episode, pinpointing the exact location and even providing images. The victim was pulled from the bottom of the pool and revived. “In this case, the AI technology improved the response time of our lifeguard and allowed him to get out there faster… the AI alert could have easily been the difference between life and death,” said Stirling’s Deputy Mayor Suzanne Migdale.
The guardian on your wrist
The devices we carry every day are becoming sophisticated safety systems that can summon help when we cannot.
Let’s keep going with the personal accounts – there’s no shortage of them, after all. Consider the account of Eric Zollinger, who probably never expected his Apple Watch (of all things) to save his life. The New York real estate broker was cycling home through Manhattan in March 2024 when he hit a flooded pothole and crashed hard onto the pavement. Dazed and injured with a fractured nose and concussion, he managed to reach his apartment but subsequently collapsed in his bathroom, all alone. Within seconds, his Apple Watch Series 8 detected the fall and his subsequent immobility, automatically dialling 911 and transmitting his exact GPS location to emergency services and designated contacts.
Paramedics arrived promptly and found Zollinger unconscious but breathing. Hospital scans revealed no life-threatening injuries, and doctors told him that without the watch’s intervention, the outcome could have been far worse. If he had been discovered the following morning, his condition would have been markedly worse. His case joins a growing catalogue of similar rescues: Apple Watches detecting strokes in Cincinnati, car crash sensors summoning help for unconscious drivers in Kenya, and heart rhythm monitors alerting users to dangerous cardiac conditions before they become critical.
The technology behind these rescues relies on sophisticated sensor arrays and machine learning algorithms. Modern smartwatches contain accelerometers and gyroscopes calibrated to distinguish between normal activity and emergencies. Apple’s fall detection system monitors for hard impacts followed by immobility, whilst crash detection uses similar principles to identify vehicle accidents. If the device detects an emergency and the user remains unresponsive for about a minute, it automatically places an emergency call and shares location data with contacts.
Keeping threats at bay
For women facing the persistent threat of violence, new AI applications are developing even more nuanced protective capabilities. Epowar, created by two University of Bath graduates, represents a breakthrough in automated threat detection. The app analyses biometric and motion data from smartwatches to detect physical attacks in real time, monitoring for distinct irregular movements and heart rate spikes that are the hallmarks of fear or distress. “Looking at movement that’s random and irregular, coupled with heart rate indicators of fear, we can very accurately tell if somebody’s being attacked,” explains co-founder E-J.
The app runs continuously in the background on devices like Apple Watch, providing automatic attack detection without requiring users to activate panic buttons during moments of crisis. When the AI identifies an assault, it triggers emergency alerts to pre-selected contacts, records evidence automatically, and compiles sensor data into secure ‘evidence packs’ that survive in the cloud should the device be destroyed. The app aims to address a real need: 81% of young women in the UK report feeling unsafe walking alone in the dark. By leveraging wearables many already use, Epowar provides silent protection that could also aid prosecutions – crucial when less than 3% of UK sexual assault cases currently result in conviction.
The technology also extends into home security, where systems like X-face.ai utilise facial recognition to identify specific individuals approaching properties. Developed specifically for domestic violence survivors, the system can recognise abusive ex-partners or stalkers from security camera feeds and immediately alert residents via secure messaging apps, providing crucial advance warning before confrontations occur.
What comes next?
What 2030’s safety landscape will look like – and what we can do about it.
By now, the trajectory of development is becoming readily apparent: we’re moving towards AI doing more than assisting in safety: it actively monitors situations to ensure it. According to a Deloitte global study, big data analytics and AI‑driven surveillance will become standard best practices in urban policing by 2030. “Introducing technology like gunshot detection empowers your police officers and law enforcement agencies to respond and help the community. There is a lot of mistrust… and traditionally marginalised low-income communities are less likely to call for help. This tech bridges that gap by sending help even when people don’t call 911,” says Jeff Merritt, Head of IoT and Urban Transformation at the World Economic Forum.
In the not-too-distant future, drones will likely become as commonplace as police cars. The FAA projects a staggering 300% increase in public safety drone fleets by decade’s end, with most mid-to-large police departments operating drone units. Many will implement ‘drone as first responder’ programmes wherein autonomous UAVs will respond to incidents before human officers are capable of arriving on the scene. DARPA, for its part, has forecast that unmanned systems could handle 30% of high-risk police tasks by 2030, from reconnaissance to bomb disposal.
Emergency response will also be transformed around AI. Consulting firm Gartner predicts that up to 50% of emergency calls will first be handled by AI bots or algorithms by 2030, which will triage and route calls more efficiently to the proper services. A white paper by Mission Critical Partners has forecast that by 2030, AI virtual assistants could handle routine 911 intakes from end to end, reducing call-processing times by 20–30% and filtering out non-emergencies automatically.
Sensors everywhere
The market is, inevitably, responding to this rising demand. The global public safety software market is expected to surge to US$36.47bn by 2030, roughly doubling its 2024 value. Meanwhile, an estimated 94 million US households (about 73% of homes) now use some type of security system in 2024. Over 50% of American homes have at least one security camera, a rate that surged significantly from the prior year. Personal safety devices represent perhaps the fastest-growing segment. Industry analysis by Future Market Insights projects global sales to grow from US$44.85bn in 2024 to US$148.23bn by 2034. These aren’t just panic buttons anymore, but intelligent systems that can detect falls, monitor vital signs, and summon help autonomously.
This technological revolution will not occur in a regulatory vacuum. The EU’s forthcoming AI Act (likely to take effect sometime in 2025) will require risk assessments and transparency for high-risk public safety AI systems. In practical terms, this means that, by 2030, any AI used in policing or emergency response in Europe must be audited for bias and accuracy. We may see mandatory ‘algorithm accountability boards’ in police departments that review how AI tools (from facial recognition to predictive crime maps) are used, with community representatives involved. Communities themselves remain divided. A 2024 poll found 56% of US adults somewhat or fully trust police to use facial recognition responsibly, while 73% believe the technology works effectively. This suggests public acceptance with reservations: people want the benefits, but remain worried about misuse.
“If you’re going to trade your privacy and freedom for security, the first question you need to ask is: Are you getting a good deal?”
Jay Stanley, senior policy analyst at ACLU
The rise of the surveillance state
As safety technology becomes more powerful and pervasive, it raises questions about privacy, bias, and control.
For all its promise, the rise of AI in public safety does raise some serious questions. Jay Stanley, senior policy analyst at the American Civil Liberties Union (ACLU), frames the dilemma thusly: “If you’re going to trade your privacy and freedom for security, the first question you need to ask is: Are you getting a good deal?” The question becomes more urgent as AI adds capabilities beyond simple observation – predicting behaviour, flagging individuals as suspicious, and making decisions that affect liberty and safety. While companies like ShotSpotter claim over 90% accuracy in gunshot detection, critics point to documented false positives that have led to unnecessary police deployments in minority neighbourhoods.
Facial recognition systems, despite improving dramatically, still exhibit higher error rates for people of colour and women, invariably resulting in wrongful arrests or harassment. The algorithms that power safety systems cannot help but reflect the biases present in their training data and programming. Predictive policing software trained on historical arrest data may perpetuate patterns of over-policing in certain communities. Automated threat detection systems might flag normal behaviour by some demographic groups as suspicious while missing genuine threats from others.
“Transparency and open dialogue are essential when it comes to implementing new technologies, especially ones that impact public safety and privacy. While developing innovative solutions like AI-driven gun detection is important… it’s equally important to ensure these initiatives are subject to public scrutiny and oversight,” argues Illinois State Rep. Kam Buckner. Yet the counterargument is compelling: these technologies are already saving lives. Remember that in Chicago alone, nearly 2,000 shooting victims received aid because of gunshot detection. Can we really afford not to use these tools? “If one life is saved with gunshot detection technology, then it is absolutely worth having,” said Chicago Alderman Ray Lopez, defending the system’s life-saving statistics during a city council debate.
Learnings
We stand at a remarkable moment in human history. The technology to dramatically reduce preventable deaths exists today, deployed in cities and homes around the world. Be it thermal drones finding lost hikers, or AI systems detecting drowning swimmers, we’ve built machines that watch over us with superhuman vigilance. The data shows these aren’t just expensive toys, but highly practical tools demonstrably saving thousands upon thousands of lives annually.
But as we embrace these digital guardians, we must remain thoughtful architects of the future we’re creating. The same AI that spots a gun in a crowded subway station could, without proper safeguards, become a tool of discrimination or malicious control. The challenge isn’t technical anymore, but moral, legal, and deeply human in its dimensions. How do we harness technology’s protective power while preserving the freedoms that make life worth protecting? The next decade will demand we answer that question not with fear or blind faith, but with wisdom, transparency, and an unwavering commitment to human dignity. Our lives, quite literally, depend on getting it right.
