Partly, and unevenly. Camera analytics and remote monitoring have already absorbed a share of what static guards used to do, which is why the Bureau of Labor Statistics projects employment of security guards to grow just 1 percent from 2025 to 2035, slower than average, adding about 11,200 positions to a base of 1,304,000. Median pay was $38,050 in 2025.
That is a very large occupation growing very slowly. Understanding which parts are shrinking and which are holding matters more here than in most jobs, because the category covers work that has almost nothing in common.
Key points
- Watching screens is the exposed part. Video analytics detects motion, loitering and perimeter breaches more consistently than a person on a night shift.
- Physical presence is the durable part. Responding, intervening, de-escalating and simply being visible cannot be done remotely.
- Employment grows 1 percent to 2035, roughly 11,200 additional positions against a base of 1.3 million.
- The work is consolidating. One operator in a monitoring centre now covers sites that once had a guard each, while response teams stay local.
- Pay is low and turnover is high, which shapes the industry more than technology does.
What security guards actually do
The occupation covers several distinct jobs that are counted together, and they face very different pressure.
Static observation. Sitting at a desk or in a control room, watching cameras and monitoring alarms. This is the most exposed category, because detection is exactly what analytics does well.
Access control. Checking credentials, signing visitors in, managing deliveries, deciding who gets through a door. Partly automated by badge systems and increasingly by facial recognition, but a person still handles the exceptions and anyone without valid credentials.
Patrol and physical presence. Walking a site, checking doors, being visible. Visibility itself is the deterrent, and a camera does not provide it in the same way.
Response and intervention. Approaching someone, asking them to leave, managing a confrontation, providing first aid, evacuating a building. This is the hardest to automate and the part with the most legal weight.
Customer-facing work. In retail, hospitals, schools and offices, a large share of a guard’s day is directions, reassurance and low-level conflict handling. Employers frequently describe this as the main reason they staff a post at all.
Anyone reading a national projection for this occupation should hold those five categories in mind. A control-room role and a hospital response post face opposite pressures, yet both are counted in the same 1.3 million. The average tells you very little about either.
Where AI is already deployed
The technology is not arriving; it has arrived, and most large sites use some form of it.
Video analytics detects motion, crossing a line, loitering, abandoned objects and crowd density, and does so continuously without fatigue. Licence plate recognition manages vehicle access. Facial recognition handles credentialed entry in some settings, subject to considerable legal variation between jurisdictions. Alarm verification uses analytics to decide whether an alert is real before dispatching anyone, which cuts false call-outs sharply. Remote monitoring centres let one operator cover many sites, with local response called only when something needs a person.
The pattern is clear. Detection is being automated. Decision, response and presence are not.
The consolidation effect is worth understanding properly, because it is where the job losses actually sit. A site that once had a guard watching its own cameras through the night now feeds those cameras to a monitoring centre. One operator there might cover twenty sites. Twenty posts become one, plus a response capability that is often shared. Nobody was replaced by a robot. The work was gathered into one place and staffed at a fraction of the previous headcount.
That is why the projection is flat rather than negative. Sites keep adding cameras and coverage, which creates monitoring work, while each individual site needs fewer people watching them.
Why the human part holds
Four things keep people in this work even as detection automates.
Someone has to respond. A camera that identifies an intruder has not removed them. Every automated detection still ends with a person going to the location, and response times are the reason clients pay for local staffing.
Liability requires a person. Detaining someone, using force, refusing entry and giving evidence in court all involve legal exposure that sits with a named, trained, licensed individual. No vendor has offered to hold that.
Deterrence is physical. A visible guard changes behaviour in a way a camera does not, which is why retailers and venues keep staffing posts that analytics could technically cover.
False positives are expensive. Analytics generates alerts that need human judgement to sort. A system that flags every fox and plastic bag has not reduced the workload unless someone competent is triaging.
There is also a failure mode that only shows up over time. When almost every alert turns out to be nothing, people stop treating alerts as meaningful. Staying alert through a long shift of false positives is genuinely difficult, and it is a known problem in monitoring work rather than a sign of poor staff. Good operations rooms design around it with rotation and workload limits; bad ones simply add cameras and wonder why response degrades.
There is a fifth factor specific to this industry. Guarding is a low-margin contract business with high turnover, and clients buy on price. That makes technology adoption uneven. Large sites with capital invest in analytics. Smaller contracts continue much as before, because the equipment costs more than the labour it would displace.
This is a general point about automation that gets missed. A technology can be entirely capable and still not be adopted, because the labour it replaces is cheap and the capital cost is not. Security guarding sits squarely in that position across much of the market, and it is the main reason the decline has been slow rather than sharp.
What to know before deciding
| Measure | Security guards, 2025 |
|---|---|
| Median annual pay | $38,050 |
| Number of jobs | 1,304,000 |
| Projected growth, 2025 to 2035 | 1 percent (Slower than average) |
| Projected employment change | 11,200 |
| Typical entry-level education | High school diploma or equivalent |
Two comparisons make the position clearer.
The occupation is enormous. At 1.3 million positions it is one of the largest in the handbook, so even flat growth means a great many openings from turnover each year. Entry remains easy relative to most work, which is both the attraction and the reason wages stay low.
The contrast with adjacent security work is stark. Information security analysts are projected to grow 21 percent with median pay of $129,180. Physical and digital security are both growing fields, but almost all of the wage and growth is on the digital side, and that gap is the single most useful fact in this article for anyone planning a career.
For context on how exposure is measured, the Bureau publishes AI exposure categories for 831 occupations and notes that exposure “does not imply job loss, productivity gains, automation probability, or wage effects.”
What actually changes over the next decade
- Monitoring consolidates. Fewer people watching screens on site, more in centralised operations rooms covering many locations.
- Response stays local. Sites keep staff for intervention and presence even when observation moves elsewhere.
- Alert triage becomes the skill. Sorting genuine events from analytics noise is now a real competency.
- Licensing requirements tighten. More jurisdictions require formal training, which raises the floor and slightly raises pay.
- The customer-service share grows. As detection automates, more of the remaining day is dealing with people.
- Robots patrol some sites. Ground units are in use in car parks and warehouses, mainly as mobile cameras rather than as responders.
- Evidence handling grows. More footage and more automated alerts mean more incidents that end up needing a written report or a court statement from a named person.
Decision framework
Five questions before taking or staying in this work.
- Which type of post is it? A control-room seat and a hospital response role are different jobs with different futures, even though both are counted as security guard.
- Is this a step or a destination? Guarding is a common entry point into policing, emergency services, facilities management and corporate security, and it works well as a route into those.
- Can you get licensed and trained beyond the minimum? First aid, conflict management, fire safety and control-room qualifications all separate you from the general pool.
- Would digital security suit you better? The pay and growth gap is very large, and the entry requirements are more about study than credentials.
- Will you learn the systems you work with? Guards who understand what the analytics is detecting, and where it produces false alerts, become supervisors rather than being displaced by the system.
That last skill transfers well beyond security. Knowing how these systems reason, and where their confident output needs checking, is general rather than sector-specific. If you want a structured route in, explore Coursiv AI lessons and check current plan details on the official site.
Your next step
If you are working in this field now, find out whether your site is heading toward remote monitoring. That single fact tells you more about your own position than any national projection, and it is usually visible a year or more ahead through equipment upgrades.
One more thing worth doing early. If guarding is a route into something else rather than the destination, decide now what that something is and check what it actually requires. Policing, emergency services, facilities management and corporate security all draw from this pool, but each has its own entry requirements, and the guards who move across are the ones who worked backwards from those requirements rather than waiting to be noticed.
Then pick one qualification beyond the licensing minimum, ideally conflict management or a control-room certification. In an occupation this large with turnover this high, small credentials move you out of the general pool faster than anywhere else.