Skip to sources
Time Machine

Security Guards

Scrub through 186years of this role's history, from when it first emerged, through every wave of technology that reshaped it, to the cited projections for where it's heading next.

2026drag to travel through time
1850187519001925195019752000now
Country
2026
Known today as Security Guards (BLS SOC 33-9032)
US Employment
1.28M
OEWS is a point-in-time survey snapshot, not a continuous time series; BLS advises against using it for year-over-year trend comparison.
Median Annual Wage
$38,020
≈ $37,045 in 2024 dollars
Each dot is a cited figure over time; the dotted line only links them (values between aren't measured). Hollow dots are estimates.
Tools of the era

The tools that defined the work

Select an era to see how it reshaped the work.

  • Nightstick + lantern + whistle (the watchman's kit)

    The foundational equipment of the private security guard in the Pinkerton era was nearly identical to that of the public police officer: a wooden nightstick or baton, a kerosene or oil lantern for night rounds, a whistle to sound alarm over a city block's distance, and a notebook for recording the post log. What distinguished the Pinkerton agent or industrial guard from the lone night watchman was organizational backing — a central office, a uniform, a chain of command, and the implicit authority of a firm with national reach and a formidable reputation. The tool was the person; the organization was the multiplier.

    Effect on the work

    The Pinkerton model scaled private security from a solo-watchman model to an organized industry for the first time. By the 1880s, the agency employed more agents than the US Army had soldiers — entirely on the strength of organizational structure, not any new technology.

    Work toolChanging equipment
  • Watchclock guard-tour system (Newman / Detex, 1878 onward, patented 1901)

    The watchclock was an ingenious mechanical accountability device: a portable clock in a leather case, worn over the shoulder like a canteen, that punched a timestamp onto a paper tape whenever the guard inserted a unique key at a checkpoint box mounted on a wall, doorframe, or pole. A night watchman making rounds through a factory, office building, or warehouse would insert the key at each station — "punching the clock" — and the device created an irrefutable paper record of the route and timing. Abraham A. Newman patented a landmark embossing mechanism in 1901; by 1902 insurance organizations had approved the system for verifying patrol logs. In 1923 the Detex Watchclock Corporation formed through merger and became the dominant manufacturer. The Guardsman model, introduced in the 1930s, could record 96 hours of patrol activity. Detex discontinued its last mechanical watchclocks in December 2011.

    Effect on the work

    The watchclock transformed security guard work from an honor-system occupation into a verifiable, auditable service. Insurance companies could now require proof of rounds; clients could see whether their overnight guard had actually walked the post or slept in the booth. It created the performance-management infrastructure of the industry a century before digital surveillance.

    Work toolChanging equipment
  • Two-way radio / walkie-talkie (post-WWII commercial deployment)

    World War II surplus radio technology entered the civilian market after 1945; Motorola leveraged its wartime experience to introduce the first transistorized portable radios in the mid-1950s. By the late 1950s, security companies were deploying handheld two-way radios to guards on patrol, enabling real-time communication with a central station for the first time. Before two-way radio, a guard on a large campus or industrial plant could go hours without communicating with anyone; the radio transformed the isolated overnight shift into a networked operation. The guard on post could now summon backup, report an incident, or be redirected from the central station in real time. This was the first technology that genuinely extended the reach of a security officer beyond the range of a whistle.

    Effect on the work

    Two-way radio improved incident response times and enabled fewer supervisors to manage larger guard forces; it also made the guard more accountable — the radio check-in replaced the physical watchclock punch as the primary presence-verification tool in many operations.

    Work toolChanging equipment
  • Analog CCTV + VCR recording (the monitor wall era)

    By the early 1970s, CCTV systems — closed-circuit television cameras connected by coaxial cable to monitors at a central security station — had become commercially viable for large commercial and industrial installations. Banks, retailers, and large office buildings began installing fixed cameras at entrances, vaults, and cash registers. The introduction of the VHS VCR in the late 1970s made continuous 24-hour recording economically practical for the first time: a single VCR with a time-lapse adaptor could record 72-240 hours onto a single tape. By the 1980s, the "security guard monitoring a bank of CCTV screens" had become the defining image of the profession in popular culture. The analog CCTV era created a new specialty within security work — the monitor operator, sitting in a control room, scanning screens — and began the long process of shifting some of the guard's situational awareness from physical patrol to electronic surveillance.

    Effect on the work

    Analog CCTV began a decades-long substitution dynamic: cameras covered positions that previously required a guard's physical presence. A single monitor operator could watch feeds from a dozen cameras simultaneously, replacing up to a dozen fixed-post guards — though the quality of pre-digital footage made real-time threat detection difficult and evidence use unreliable.

    Bedside monitoringVitals at a glance
  • IP camera + digital video recorder + access-control integration (the DVR era)

    The transition from analog CCTV to IP-based network cameras and digital video recorders (DVRs) through the 2000s transformed surveillance from a tape-archive function into a searchable, networked intelligence system. IP cameras transmitted over standard Ethernet; footage was stored on hard drives rather than VHS tapes; remote viewing from a laptop or smartphone became possible. Access-control systems — card readers, biometric scanners, electronic door locks — integrated with the video layer, creating the first unified physical security platforms. A single security operations center could now monitor multiple buildings or campuses from one location. The manpower required for routine camera monitoring fell; the scope of what a single guard could oversee from a desk expanded dramatically.

    Effect on the work

    IP camera deployments with DVR archival began significantly reducing the number of fixed-post guard positions at camera-saturated installations like retail chains, parking garages, and corporate campuses. Staffing models shifted toward fewer, more technically capable operators managing broader surveillance footprints.

    Work toolChanging equipment
  • AI cloud camera analytics — Verkada, Eagle Eye Networks (the intelligent-eye era)

    Verkada, founded in 2016 by three Stanford graduates and former Meraki co-founder Hans Robertson, combined cloud-managed security cameras with on-board AI analytics: object detection, person recognition, license-plate reading, and behavioral analysis running at the edge. A Verkada installation required no on-premise server; footage streamed to the cloud and was accessible from any browser. By 2022, Verkada had raised $205 million at a $3.2 billion valuation; by 2025 it had raised more than $700 million at a $5.8 billion valuation. Eagle Eye Networks (founded 2012) built a similar cloud-video-surveillance platform targeted at commercial chains and multi-site enterprises. The AI analytics layer — which could automatically flag a person loitering near a restricted area, detect an unattended bag, or alert when a car's license plate matched a watchlist — began replacing the monitor-operator's attentive gaze with an algorithm that never blinked.

    Effect on the work

    AI camera analytics compressed the monitor-operator role: where a human had previously needed to watch twelve screens, an AI system could watch five hundred feeds and alert on anomalies automatically. Early adopter reports suggested 30-50% reductions in control-room staffing at camera-heavy installations, though validated third-party studies are sparse.

    Work toolChanging equipment
  • Knightscope K5 + Cobalt Robotics — autonomous indoor/outdoor patrol robots

    Two companies, founded within three years of each other, brought autonomous patrol robots from research to commercial deployment. Knightscope (founded 2013, IPO January 2022) built the K5, a five-foot, 400-pound outdoor patrol robot on wheels that used cameras, thermal imaging, and a microphone array to monitor parking lots, campuses, and outdoor facilities. The K5 was leased to clients at a reported rate of $7-9 per hour — below the minimum wage for a human guard. Cobalt Robotics (founded 2016 by SpaceX and Google X alumni) built indoor patrol robots deployed in office lobbies and corporate campuses, with 60+ sensors and a remote human operator available for video interaction. Cobalt had raised over $90 million and deployed hundreds of robots at FedEx, DoorDash, Yelp, and other Fortune 500 companies by 2022. Knightscope reported over 10,000 machines deployed by its IPO period. Neither replaced guards wholesale: both companies positioned the robots as force multipliers, with human guards responding to alerts generated by the machines.

    Effect on the work

    At installations where Knightscope K5 or Cobalt robots were deployed, clients typically reported reducing fixed-post guard hours by 20-30% at the specific patrol area. The robots did not eliminate guards from sites; they eliminated the routine walk-around shifts, leaving human guards for access control, escalation, and customer interaction.

    Work toolChanging equipment
  • AI gunshot detection + loitering detection + Flock Safety LPR + facial recognition (pervasive AI sensing)

    By 2020, the AI analytics layer in private security had expanded far beyond camera-monitoring into a suite of autonomous sensing capabilities. Flock Safety (founded 2017), which began as a neighborhood license-plate reader network for law enforcement, had expanded by 2025 to operate in over 5,000 communities across 49 states performing over 20 billion vehicle scans per month — technology that private security operations now integrated into campus perimeter protection. AI loitering detection (a person remaining stationary in a restricted zone for more than a configurable time threshold) and AI gunshot/sound-anomaly detection became standard features of commercial security platforms. Facial recognition databases, controversially, began integrating with access control at some corporate and retail installations. The cumulative effect of these capabilities is that a modern security operations center with AI tools can monitor a physical campus with the situational awareness that previously required two to three times the guard headcount — but someone still has to respond when the alarm fires.

    Effect on the work

    The AI-sensing era has not eliminated the security guard; it has changed the ratio between human response officers and electronic sensors. Large campuses that previously staffed ten patrol guards per shift now staff three or four response officers backed by an AI-monitored camera and sensor network. The remaining humans are harder to automate: they speak to distressed employees, de-escalate the person at the front desk who cannot get access to an office, and make the physical-presence decision that no camera can make.

    Work toolChanging equipment
Projection cone · present → 2034

What credible sources project

Scrub the slider past now to anchor each scenario on the scrubber. The spread is the range of futures credible sources project for this role.

Employment outlook
Projected change in the number of people doing this work.
Presence-as-deterrent floor scenario (human escalation cannot automate)
2033
+5%
Optimistic scenario grounded in the irreducible human functions in security: physical deterrence by visible presence (a uniformed person is a different deterrent than a camera), legal authority (most jurisdictions require a licensed human officer to make a citizen's arrest or detain an individual), customer service at commercial and hospitality venues, and escalation judgment in ambiguous situations. If AI cameras and robots handle the routine surveillance and patrol functions while human guard deployment concentrates in the presence-as-deterrent and escalation roles, headcount growth is possible as the broader category of "security-related work" expands with AI monitoring of more spaces. This is the bull case: more premises monitored by AI, staffed with smaller human response teams, resulting in a slight net growth in security personnel as total monitored premises expands faster than AI displaces positions per site.
BLS Occupational Outlook 2023-33
2033
+2%
BLS Employment Projections 2023-33 cycle. Published employment change for security guards and gambling surveillance officers as a group: approximately +2%, described as "slower than the average for all occupations." Despite modest employment growth, BLS projects approximately 162,400 annual openings — a large number driven almost entirely by replacement need (turnover in this occupation is among the highest of any service occupation). The +2% net-employment figure reflects modest demand growth offset by technology substitution and ongoing consolidation of the contract-security market. BLS explicitly cites continued demand from commercial real estate, healthcare, schools, and transportation facilities as growth drivers, against the backdrop of AI cameras and remote-monitoring tools reducing the need for fixed-post guards.
AI-augmentation scenario (cameras replace posts, robots replace patrols)
2034
-15%
Scenario analysis based on documented patrol-robot deployments (Knightscope, Cobalt), AI camera analytics adoption (Verkada, Eagle Eye), and the structural economics of the industry. If AI cameras and patrol robots continue to penetrate the fixed-post and routine-patrol segments of the market — historically the largest components of guard headcount — while the response-officer and customer-facing segments remain human, the net employment trajectory could decline 10-20% from the 2023 baseline by 2034. The pessimistic end of this scenario assumes that: (a) Knightscope-style outdoor robots become cost-competitive with minimum-wage guard labor in 50%+ of outdoor patrol applications; (b) AI camera analytics fully substitute for monitor-room headcount in commercial installations; and (c) the remaining response-and-escalation workforce is 60-70% of the 2023 level. The +2% BLS figure is the optimistic anchor; the -15% scenario is the pessimistic tail if hardware costs for autonomous patrol continue to fall.
AI task exposure
Share of the role’s tasks that researchers estimate AI can do. This is a measure of task exposure, not a forecast of jobs lost.
Frey & Osborne (2013)
2033
35%
of tasks
Gaussian-process classifier on O*NET task features. Frey & Osborne assigned Security Guards a probability of computerization of approximately 0.84 — placing them in the high-risk tier of the 702-occupation dataset. The bottleneck factors they identified were relatively weak: security-guard tasks involve some social intelligence (responding to people, de-escalation) but are not primarily social; they involve manual mobility but mostly routine patrolling rather than the complex manipulation that protects electricians or surgeons. The 0.84 score reflects F&O's assessment that autonomous patrol (robots), AI surveillance (cameras), and remote monitoring could substitute for the routine physical-presence functions that dominate guard hours. The -35% figure represents a conservative displacement below the full F&O probability if realized over the decade. In practice, employment has remained near 1 million through 2024 — suggesting F&O's timeline for automation penetration was optimistic — though Knightscope and Cobalt robot deployments validate the direction of the finding.
Eloundou et al. — 'GPTs are GPTs' (2023)
2028
3%
of tasks
GPT-4 task-by-task LLM exposure labeling on O*NET tasks. Security guards score very low on LLM exposure: the core tasks — patrolling premises, monitoring CCTV, checking credentials, responding to alarms, detaining suspects — are not text-based work that a language model can perform. The occupation's high Frey & Osborne automation risk comes from physical robotics, not LLMs. The -3% estimate represents the conservative near-term impact from AI-assisted tools (AI analytics platforms reducing monitor-operator headcount) rather than from robotics-driven physical patrol displacement. Eloundou et al. classify security guards in the low-LLM-exposure tier; the real risk for this occupation comes from non-LLM AI (computer vision, autonomous mobile robots) that the Eloundou framework does not score.
Today, in this role

What's shifting in the work right now

The historical view above shows how this role has moved. This is the present-day detail: which AI tools are picking up which tasks, where the edge still is, and the natural directions this work can grow.

What's changing in your day

Three parts of your work where AI is already doing real lifting, and what stays yours.

AI is sitting alongside you hereWrite incident reports documenting occurrences during the shift — recording the time, location, parties involved, sequence of events, actions taken, and disposition of each incident, and submitting the completed report through an electronic security management platform for supervisor review and client records.

Write incident reports documenting occurrences during the shift — recording the time, location, parties involved, sequence of events, actions taken, and disposition of each incident, and submitting the completed report through an electronic security management platform for supervisor review and client records.[1],[11]

Where your edge is

Incident reporting is the task most exposed to AI assistance in the near term. Digital Guard Tour and platforms like TrackTik already auto-populate the time, location, checkpoint scans, and responding guard identity from GPS and NFC data — the guard only needs to describe what happened and what they did. Some platforms are beginning to use voice-to-text capture during the incident itself to reduce post-shift report writing time. The discipline that remains irreplaceable: describing the facts accurately in plain language, capturing the details that matter legally (exact time of contact, exact words used, whether the subject complied voluntarily), and flagging when an incident may have liability implications for the client. Write the report like a police officer would: who, what, where, when, how — no adjectives, no speculation, just documented observation.

AI is sitting alongside you hereMonitor AI-assisted video surveillance systems — watching live camera feeds aggregated in a video management system (VMS), reviewing AI-generated alerts for motion anomalies, loitering, perimeter breaches, or access violations, confirming true alarms and dismissing false positives, and dispatching a physical response or notifying emergency services when the alert is verified.

Monitor AI-assisted video surveillance systems — watching live camera feeds aggregated in a video management system (VMS), reviewing AI-generated alerts for motion anomalies, loitering, perimeter breaches, or access violations, confirming true alarms and dismissing false positives, and dispatching a physical response or notifying emergency services when the alert is verified.[1],[9],[12]

Where your edge is

Camera monitoring is the task most transformed by AI right now — and the biggest opportunity to move up. Verkada, Avigilon Unity, and Scylla AI filter false positives with 99%+ accuracy, which means the guards who can confidently use these platforms spend almost no time on phantom alarms and respond only to verified threats. The skill that separates a $18/hr monitor from a $28/hr remote operations specialist is the ability to configure AI alert thresholds, understand why the system flagged an event, and make rapid verification decisions. Learn your VMS: what 'loitering' is configured to mean on your site, how Avigilon's Appearance Search works, and how to pull a multi-camera timeline for an incident investigation. Those skills are the same ones security directors say they cannot find enough of in 2025.

AI is sitting alongside you hereOperate as a remote monitoring officer — from a central station, monitoring camera feeds for multiple sites simultaneously using AI-assisted VMS dashboards, triaging AI-generated alerts, issuing verbal deterrence through site speakers, dispatching mobile patrol units or emergency services to confirmed incidents, and maintaining a shift log of all monitored events.

Operate as a remote monitoring officer — from a central station, monitoring camera feeds for multiple sites simultaneously using AI-assisted VMS dashboards, triaging AI-generated alerts, issuing verbal deterrence through site speakers, dispatching mobile patrol units or emergency services to confirmed incidents, and maintaining a shift log of all monitored events.[4],[13],[5]

Where your edge is

Remote monitoring officer is the highest-growth and highest-pay variant of the security guard role in 2025-2026. ASIS data shows AI-literate guards who can monitor 10-30 sites from a central operations center in higher demand than any other guard category. The role requires comfort operating multiple software interfaces simultaneously, the ability to triage AI alerts under time pressure, and clear verbal communication over speaker systems to deter threats at a distance. Pay is typically 20-40% above on-site guard rates for the same experience level, and the schedule (mostly indoor, no physical patrol) attracts guards who want longevity without the physical attrition. The fastest path to this role: get certified on Verkada, Avigilon, or Genetec platforms (vendor training programs exist for all three), and volunteer for any remote monitoring assignments your current employer offers.

Where this role is heading

Natural next steps for someone with your foundation: not exits, evolutions.

A direction you could grow

Security Managers

Security Manager (SOC 11-3013.01) is the natural promotion trajectory for an experienced guard or security officer — typically requiring 5-10 years of field experience, a demonstrated track record of competent incident management, and increasingly a certification such as CPP (Certified Protection Professional) or PSP (Physical Security Professional) from ASIS International. Security managers oversee guard forces (often 10-50+ officers), manage vendor relationships for camera systems and access control, develop site security plans, conduct threat assessments, and interact directly with C-suite leadership on physical security strategy. The AI tools that augment frontline guard work (Verkada, Avigilon, Hakimo) are the same tools a security manager must be able to evaluate, procure, configure, and supervise — so guards who develop technical fluency with these platforms now are building exactly the skills that security director roles require in 2025-2026. BLS OEWS 2024 median for Security Managers: $109,990/yr, nearly triple the security guard median of $38,370.

What you'd add
· ASIS Certified Protection Professional (CPP) or Physical Security Professional (PSP) certification
· Security operations management: guard scheduling, post orders, performance evaluation, progressive discipline
· Physical security technology: AI VMS configuration, access control system administration, alarm system management
· Threat and vulnerability assessment methodology: site surveys, risk registers, security plans
· Budget management and vendor contract negotiation for security services and equipment
What it takesSome new skills to pick up
Share this year
Drops anyone you send it to straight into 2026.
Preview card
Part of Protective Services · see all 14roles →
Different role?

See the same long-arc view for your own profession.

Browse the directory by industry, or search by title or SOC code. New roles ship every few weeks. Every profile cites every claim.

Browse all roles

The data behind this timeline

On record since1850
Latest tracked employment1,283,470 (US, 2025)
Latest median pay$38,020 (2025)
Outlook+2% by 2033 (BLS Occupational Outlook 2023-33)
View all 28 cited data points
YearUS employmentMedian annual paySource
190035,000n/aESTIMATE
1940100,000n/aESTIMATE
1970400,000n/aESTIMATE
1985750,000$11,500ESTIMATE
20001,065,000$18,500BLS-OEWS
2003964,260$19,660BLS-OEWS
2004978,570$20,320BLS-OEWS
2005990,000$20,760BLS-OEWS
20061,004,130$21,530BLS-OEWS
20071,032,260$22,570BLS-OEWS
20081,046,760$23,460BLS-OEWS
20091,028,830$23,820BLS-OEWS
20101,047,000$23,920ESTIMATE, BLS-OEWS
20111,032,940$23,900BLS-OEWS
20121,046,420$23,970BLS-OEWS
20131,066,730$24,070BLS-OEWS
20141,077,520$24,410BLS-OEWS
20151,097,660$24,630BLS-OEWS
20161,103,120$25,770BLS-OEWS
20171,105,440$26,900BLS-OEWS
20181,114,380$28,490BLS-OEWS
20191,130,000$30,750BLS-OEWS
20201,054,400$31,050BLS-OEWS
20211,057,100$31,470BLS-OEWS
20221,124,890$34,750BLS-OEWS
20231,200,000$36,470BLS-OEWS
20241,241,770$38,370BLS-OEWS
20251,283,470$38,020BLS-OEWS
Embed this timeline on your site

Free for any site. Paste this where the timeline should appear; it stays interactive, every datapoint stays cited, and it sets no cookies on your page. How embedding works

<iframe src="https://futurehistory.earth/embed/33-9032"
  width="100%" height="430" style="border:0"
  title="Security Guards, a Future History timeline"
  loading="lazy"></iframe>

See all roles in Protective Services