Transit and Railroad Police

O*NET-SOC code: 33-3052.00

Protect and police railroad and transit property, employees, or passengers.

Data source: BLSPublished: 2026-08
Moderate· relative
LowFour relative bandsVery high
Group-level value
Scale, basis and source
Four relative bands (Low / Moderate / High / Very high)
831 detailed occupations in the BLS Employment Projections table. Assigned per National Employment Matrix (NEM) code, so occupations sharing a NEM code carry the same band
Source dataset (XLSX download)
Data source: AnthropicPublished: 2026-03
0.000
0.000Range of values carried here0.745
Scale, basis and source
Observed exposure index, 0–1 as published
Mapped onto O*NET tasks
Published per SOC 2018 occupation; every O*NET occupation with the same SOC 2018 code carries this value
Source dataset (CSV download)
Data source: ILOPublished: 2025
0.14
0.09Range of values carried here0.70
Group-level value
Scale, basis and source
Generative AI exposure index, 0–1 as published
Published per ISCO-08 unit group. Linked to this occupation, wholly or in part, by applying U.S. Bureau of Labor Statistics crosswalks (ISCO-08 to 2010 SOC, 2010 SOC to 2018 SOC) as published
Source dataset (PDF download)

What kind of figure this source publishes

The BLS category is a relative rank, not an absolute level, and it is not a first-hand measurement: it groups an occupation's percentile ranks across several published studies into four bands. It is not an employment or wage forecast, not a probability of adoption, and it does not separate automation from augmentation.

Task-level exposure

14 rated tasks · 6 tasks with β ≥ 0.5 (42.9%)

β = direct exposure (E1) + 0.5 × exposure when tools are available (E2), per the source repository's definition.

Source unit: O*NET 27.2 tasks → O*NET 31.0 occupation code
2 rated tasks are not in the O*NET 31.0 task list; their scores are kept as published for O*NET 27.2.

Source: OpenAI "GPTs are GPTs" exposure rubric
Release: gh-main-0471612
License: MIT License, Copyright (c) 2024 OpenAI

Task (O*NET 30.2)	Claude.aiRaw / share %	APIRaw / share %
Prepare reports documenting investigation activities and results. 33-3052.00	0.00000.0	—0
Examine credentials of unauthorized persons attempting to enter secured areas.	—0	0.00000.0
Not observed on any surface — 10 task(s) — These tasks have no row in the source for this release. The 0 in the share row is a display-stage composition ratio; absence is what the — in the raw row states.
Patrol railroad yards, cars, stations, or other facilities to protect company property or shipments and to maintain order.	—0	—0
Apprehend or remove trespassers or thieves from railroad property or coordinate with law enforcement agencies in apprehensions and removals.	—0	—0
Investigate or direct investigations of freight theft, suspicious damage or loss of passengers' valuables, or other crimes on railroad property.	—0	—0
Direct security activities at derailments, fires, floods, or strikes involving railroad property.	—0	—0
Plan or implement special safety or preventive programs, such as fire or accident prevention.	—0	—0
Monitor transit areas and conduct security checks to protect railroad properties, patrons, and employees.	—0	—0
Enforce traffic laws regarding the transit system and reprimand individuals who violate them.	—0	—0
Provide training to the public or law enforcement personnel in railroad safety or security.	—0	—0
Direct or coordinate the daily activities or training of security staff.	—0	—0
Interview neighbors, associates, or former employers of job applicants to verify personal references or to obtain work history data.	—0	—0

Values in this tab are predicted labels, not observations. Eloundou et al. (2023) published two rating regimes — human raters and GPT-4 — and the β shown here is derived from the GPT-4 rater basis alone; the same task can take a different value under the other regime. The unit and the meaning differ from the observed shares (%) in the other tabs, so do not place them on the same axis.

Task (O*NET 27.2)	βE1 + 0.5 × E2
Prepare reports documenting investigation activities and results. O*NET Task ID 9478	1.0
Examine credentials of unauthorized persons attempting to enter secured areas. O*NET Task ID 9476	0.5
Investigate or direct investigations of freight theft, suspicious damage or loss of passengers' valuables, or other crimes on railroad property. O*NET Task ID 9479	0.5
Interview neighbors, associates, or former employers of job applicants to verify personal references or to obtain work history data. O*NET Task ID 9482	0.5
Record and verify seal numbers from boxcars containing frequently pilfered items, such as cigarettes or liquor, to detect tampering. O*NET Task ID 9483	0.5
Plan or implement special safety or preventive programs, such as fire or accident prevention. O*NET Task ID 9484	0.5
Patrol railroad yards, cars, stations, or other facilities to protect company property or shipments and to maintain order. O*NET Task ID 9475	0.0
Apprehend or remove trespassers or thieves from railroad property or coordinate with law enforcement agencies in apprehensions and removals. O*NET Task ID 9477	0.0
Direct security activities at derailments, fires, floods, or strikes involving railroad property. O*NET Task ID 9480	0.0
Direct or coordinate the daily activities or training of security staff. O*NET Task ID 9481	0.0
Seal empty boxcars by twisting nails in door hasps, using nail twisters. O*NET Task ID 9485	0.0
Monitor transit areas and conduct security checks to protect railroad properties, patrons, and employees. O*NET Task ID 21002	0.0
Enforce traffic laws regarding the transit system and reprimand individuals who violate them. O*NET Task ID 21003	0.0
Provide training to the public or law enforcement personnel in railroad safety or security. O*NET Task ID 21004	0.0

β = E1 + 0.5 × E2 · E1 = tasks where direct LLM access alone cuts time by at least 50%, E2 = tasks where software built on top of an LLM cuts time by at least 50%. Values take only 0 / 0.5 / 1.0.

Data sources & licenses — O*NET®, Anthropic Economic Index, Eloundou et al. (2023): see full notices on the Credits page

All O*NET 31.0 task statements (12)

All O*NET 31.0 task statements for this occupation (English, as published; task type core or supplemental). Exposure values are not listed here; they are in the table above, next to each task statement as worded in the O*NET version each source used.

Task (O*NET 31.0)	Type
Patrol railroad yards, cars, stations, or other facilities to protect company property or shipments and to maintain order.	Core
Examine credentials of unauthorized persons attempting to enter secured areas.	Core
Apprehend or remove trespassers or thieves from railroad property or coordinate with law enforcement agencies in apprehensions and removals.	Core
Prepare reports documenting investigation activities and results.	Core
Investigate or direct investigations of freight theft, suspicious damage or loss of passengers' valuables, or other crimes on railroad property.	Core
Direct security activities at derailments, fires, floods, or strikes involving railroad property.	Core
Plan or implement special safety or preventive programs, such as fire or accident prevention.	Core
Monitor transit areas and conduct security checks to protect railroad properties, patrons, and employees.	Core
Enforce traffic laws regarding the transit system and reprimand individuals who violate them.	Core
Provide training to the public or law enforcement personnel in railroad safety or security.	Core
Direct or coordinate the daily activities or training of security staff.	Supplemental
Interview neighbors, associates, or former employers of job applicants to verify personal references or to obtain work history data.	Supplemental

Anthropic, "What work can robots do?"2026-09

0.81/ 3· Robot exposure index

0 none3 unstructured settings

Scale, basis and source

Each physical task is graded 0–3 by the setting in which robots that have been deployed, sold or demonstrated could do it (0 none · 1 purpose-built robot settings such as a fenced machine cell or a bottling line · 2 structured facilities such as a hospital pharmacy or a container port · 3 unstructured settings such as a private home or a construction site), and the grades are averaged by share of working time

Physical tasks among this occupation's O*NET 29.3 tasks

Anthropic estimated the grades and time shares with Claude Opus 5 using web search, as of 2026.

Source dataset

The robot exposure index is a different measure from the AI exposure figures on this site and cannot be added to or compared with them. It does not measure actual adoption or cost, and 0 also covers occupations with no physical tasks.

Tier of each physical task

Robot environment tiers (E0–E3)

E0 noneDetails
Rated not doable by robots that have been deployed, sold or demonstrated, even in a purpose-built setting
E1 purpose-built robot settingsDetails
Rated doable in a setting built around the machine, such as a fenced machine cell or a bottling line
E2 structured facilitiesDetails
Rated doable in a human work facility that an institution keeps to a working standard, such as a hospital pharmacy or a container port
E3 unstructured settingsDetails
Rated doable in a setting that no institution keeps to a working standard, such as a private home or a construction site

Only tasks classified as physical are listed; other tasks have no tier in the source.

Task (O*NET 29.3)	Robot environment tier
Patrol railroad yards, cars, stations, or other facilities to protect company property or shipments and to maintain order.	E2 structured facilities
Apprehend or remove trespassers or thieves from railroad property or coordinate with law enforcement agencies in apprehensions and removals.	E0 none
Investigate or direct investigations of freight theft, suspicious damage or loss of passengers' valuables, or other crimes on railroad property.	E0 none
Monitor transit areas and conduct security checks to protect railroad properties, patrons, and employees.	E1 purpose-built robot settings
Enforce traffic laws regarding the transit system and reprimand individuals who violate them.	E0 none

Employment outlook (BLS)

Employment, 2025: 4K
Projected employment, 2035: 5K
Change (2025–35): +3.2%
Median annual wage, 2025: $90,230
Typical education needed for entry: High school diploma or equivalent

lowest $30,890median wage across 825 detailed occupations, BLS EP 2025–35highest $559,030

Source: U.S. Bureau of Labor Statistics, Employment Projections 2025–35 — Employment by detailed occupation, 2025 and projected 2035

Korea classification links (KSCO)

Links from the current release of the Korea Occupation Direct Crosswalk v1.1, made by AI Changing Work. Relation values are read from the O*NET-SOC occupation side.

KSCO code	KSCO title	Relation	Status
41112	일반 경찰관 및 수사관	This occupation is narrower	confirmed

Relation and status definitions(kr-direct v1.1)

ILO occupational description

AI Changing Work links the ISCO-08 unit groups below to this occupation by applying the U.S. Bureau of Labor Statistics crosswalks (2018 SOC to 2010 SOC, 2010 SOC to ISCO-08) as published, without case-by-case selection; each link may be whole or partial.

ISCO-08 5412Police Officers

The definition, tasks and examples below are the ILO's description of that unit group, not of this occupation.

Police officers maintain law and order, patrolling public areas, enforcing laws and regulations and arresting suspected offenders.

Tasks

(a) patrolling a specific area to maintain public order, responding to emergencies, protecting people and property and enforcing laws and regulations;
(b) identifying, pursuing and arresting suspects and perpetrators of criminal acts;
(c) directing traffic and assuming authority in the event of accidents;
(d) providing emergency assistance to victims of accidents, crimes and natural disasters.

Example occupations

Constable · Police officer · Police patrol officer

Sources: ILO ISCO-08 — details: Credits

ESCO occupational vocabulary

Each entry below is matched to this occupation in the published ESCO-O*NET Crosswalk (version 1, 2022). The match types are that crosswalk's; the wording is AI Changing Work's.

civil enforcement officerESCO 5419.2This occupation is narrower
Civil enforcement officers patrol the streets to enforce parking restrictions, ensure a free flow of traffic is maintained, ensure the safety of pedestrians, and ensure traffic and parking laws are complied with. They aid with the combating of crime and crime reduction by providing assistance during incidents and providing patrol services.
police officerESCO 5412.1This occupation is narrower
Police officers use investigation methods to prevent crime, to pursue and apprehend criminals, and protect the public from violent and criminal activities. They perform surveillance and patrol activities to prevent crimes and apprehend offenders. They provide support to victimised parties and the public at large, and fulfil administrative duties.

Source: European Commission, DG EMPL — ESCO API 2026-08-08 (bridge v1.2.1)

License: CC BY 4.0 (EU Reuse Notice)

View original

ESCO reuse terms and the modification notice are in Credits.

France classification links (ROME)

AI Changing Work's links between this occupation and ROME 4.0 job profiles, judged with AI models. Relations are read from this occupation's side. Links marked “provisional” were adopted on the judgments of two judges from the same model family (Claude) alone, or by an operator rule applied to their judgments, so they rest on weaker evidence than confirmed links.

ROME code	ROME title (French)	Relation
K2511	Agent / Agente de sûreté ferroviaire	Same scopeprovisional
K1708	Gardien / Gardienne de la paix	Related, not a matchprovisional

Source: France Travail — Répertoire Opérationnel des Métiers et des Emplois (ROME 4.0), version 61, last updated 15 June 2026 (https://www.data.gouv.fr/datasets/repertoire-operationnel-des-metiers-et-des-emplois-rome). Licence Ouverte 2.0.

ROME codes and titles are reproduced unchanged; the links to O*NET-SOC occupations are AI Changing Work's own judgment and are not a France Travail correspondence.

(AI Changing Work link release: fr-direct v1)

Spain classification links (CNO-11)

AI Changing Work's links between this occupation and CNO-11 occupations, judged with AI models. Relations are read from this occupation's side.

CNO-11 code	CNO-11 title (Spanish)	Relation
5921	Policías nacionales	Related, not a match
5923	Policías locales	Related, not a match
5941	Vigilantes de seguridad y similares habilitados para ir armados	Related, not a match

Source: Instituto Nacional de Estadística (INE), National Classification of Occupations 2011 (CNO-11, Royal Decree 1591/2010) — www.ine.es · Licensed CC BY 4.0. CNO-11 codes and Spanish titles are reproduced unchanged; the links to O*NET occupations are AI Changing Work's own judgment and are not an INE correspondence.

(AI Changing Work link release: es-direct v1)

Germany classification links (KldB)

AI Changing Work's links between this occupation and KldB occupations, judged with AI models. Relations are read from this occupation's side.

KldB code	KldB title (German)	Relation
53212	Berufe im Polizeivollzugsdienst - fachlich ausgerichtete Tätigkeiten	This occupation is narrower
53213	Berufe im Polizeivollzugsdienst - komplexe Spezialistentätigkeiten	Related, not a match

Source: © Statistik der Bundesagentur für Arbeit — Klassifikation der Berufe 2010 – überarbeitete Fassung 2020, https://statistik.arbeitsagentur.de. KldB codes and titles are reproduced unchanged; the links to O*NET occupations are AI Changing Work's own mapping (eigene Zuordnung) and are not a Bundesagentur für Arbeit correspondence.

(AI Changing Work link release: de-direct v1)

Japan classification links (JSCO)

AI Changing Work's links between this occupation and JSCO occupations, judged with AI models. Relations are read from this occupation's side.

JSCO code	JSCO title (Japanese)	Relation
441	警察官	This occupation is narrowerprovisional
453	警備員	Related, not a matchprovisional

Source: Ministry of Internal Affairs and Communications (MIC), Japan Standard Occupational Classification (Dec. 2009, MIC Notification No. 555). JSCO codes and Japanese titles are reproduced unchanged; the links to O*NET occupations are AI Changing Work's own judgment and are not a MIC correspondence.

(AI Changing Work link release: jp-direct v1)

Sources, licenses and AI Changing Work's modification and translation notices for this page: Credits & sources