Artificial Intelligence Across Rail Networks
Loading results...
Deployment date records the documented start of use, including pilots. Source published is the article date. Source last accessed records a successful editorial retrieval, not confirmation of current deployment. Unknown dates are shown explicitly; dates use YYYY-MM-DD, YYYY-MM or YYYY.
| Country | Organization | Category | AI Application | Source & dates |
|---|---|---|---|---|
| Germany | Siemens Mobility / Rhein-Ruhr-Express (RRX) | Predictive Maintenance |
Operational depot application: Railigent X analyzes vehicle data with AI-based algorithms to predict faults and recommend preventive maintenance at Siemens Mobility's Dortmund service center, which maintains RRX trains. The 2024 source reports existing use; the planned depot expansion is a separate project.
|
View source
|
| India | Indian Railways / Northeast Frontier Railway | Safety and Surveillance |
Operational wildlife warning: an AI-enabled intrusion-detection system analyzes distributed acoustic sensing data from optical fibre to detect elephant movements near tracks. It alerts locomotive crews, station masters and control rooms. The Ministry of Railways reports operation over 141 route-kilometres.
|
View source
|
| Japan | JR East | Safety and Surveillance |
Operational safety support: AI analyzes train-mounted boarding-camera video and alerts drivers to people approaching the train. JR East reports initial use on some Sagami Line trainsets in April 2026, followed by rollout to all trainsets on that line.
|
View source
|
| Japan | JR East | Asset and Infrastructure Monitoring |
Operational decision support: JR East introduced AI assistance for dispatchers handling disruptions affecting selected conventional-line signalling equipment in the Tokyo metropolitan area in March 2023. The cited section describes a generative-AI enhancement as planned; completion of that enhancement is not established here.
|
View source
|
| Japan | JR East | Customer Experience |
Trial deployment: JR East is testing AI-enabled conversational reserved-seat ticket machines at Machida station to assist with straightforward Shinkansen ticket purchases. Its technology page also describes earlier conversational ticket-support demonstrations at Tachikawa and Omiya stations.
|
View source
|
| UK | Hitachi Rail | Predictive Maintenance |
Reported fleet-management deployment: Hitachi Rail identifies hundreds of UK trains as an HMAX success case. The AI-powered asset-management platform combines sensor data and analytics for condition monitoring and predictive maintenance. The source does not specify UK operators, a fleet-level rollout date or which optional functions each fleet uses.
|
View source
|
| South Korea | Hyundai Rotem / GTX-A | Predictive Maintenance |
Reported operational use: GTX-A condition-based maintenance collects data from 21 component types. Machine-learning and deep-learning models classify and predict time-series faults, supporting condition diagnosis, alarms and maintenance planning.
|
View source
|
| South Korea | Hyundai Rotem / GTX-A | Asset and Infrastructure Monitoring |
Reported operational use: onboard cameras and LiDAR inspect track and overhead contact-line equipment during train operation. Deep-learning image analysis detects infrastructure defects and obstacles, with alerts transmitted to the control center.
|
View source
|
| South Korea | Seoul Metro | Customer Experience |
Documented pilot: AI speech translation helps passengers and station staff communicate through a screen in 13 languages. The Myeongdong station pilot began on 4 December 2023. The March 2024 source announces expansion to 11 stations but does not establish completion.
|
View source
|
| South Korea | KORAIL | Customer Experience |
Reported service: voice AI extracts journey details from natural-language ticket requests and asks for missing information. The initial service launched in February 2025; a natural-language upgrade was reported in March 2026. App access is restricted to eligible visually or physically disabled members; telephone voice booking serves all users.
|
View source
|
| South Korea | KORAIL | General Office Work |
Reported internal deployment: AI-RailFinder uses generative AI to search, summarize and analyze railway manuals, reports and regulations. Staff can select retrieval-augmented or cache-augmented generation. Source references and department-level access permissions support controlled use of internal knowledge.
|
View source
|
| South Korea | Busan Transportation Corporation / Konan Technology | Customer Experience |
Reported deployment at Busan metro station: LingoX provides AI voice and chat assistance in 13 languages for routes, fares, tickets and nearby attractions. Complex questions transfer to a remote employee by video, with live translation supporting the conversation.
|
View source
|
| South Korea | Incheon Transit Corporation / A2Mind | Asset and Infrastructure Monitoring |
Ongoing field testing: an autonomous or remotely operated track-inspection robot gathers rail-condition data while deep-learning models are developed for fasteners and ballast or sleeper cracks. Trials on lines 1, 2 and 7 began in August 2026. Regular operation is planned for May 2027.
|
View source
|
| Thailand | Korea Railroad Research Institute (KRRI) / LOBSE / Siemens Mobility | Asset and Infrastructure Monitoring |
Field demonstration: the Korean RAILOBSE mobile inspection system demonstrated AI-assisted rail-wear and surface-defect inspection on Bangkok BTS tracks around Asia Pacific Rail in May 2026. Technology transfer and commercialization are reported; the demonstration and cooperation letter do not establish a regular fleet-wide deployment.
|
View source
|
| China | Beijing Infrastructure Investment (BII) | General Office Work |
Reported deployment: a large-language-model agent using DeepSeek and Qwen classifies and routes citizen-complaint work orders. Training uses historical cases to match incoming requests to the responsible organization and support administrative handling.
|
View source
|
| China | Beijing Rail Transit Network Administration (BII) | Scheduling and Optimization |
Reported dispatcher support: a large-language-model interface answers natural-language questions about metro passenger flows and retrieves operational data for analysis and reports. Examples include peak-period changes and passenger loads on lines 1 and 2. This is analytics support, not evidence of autonomous train control.
|
View source
|
| China | China Academy of Railway Sciences / Beijing Jingwei / Nanchang West and East Stations | Safety and Surveillance |
Reported deployment for the January 2025 Spring Festival travel period: convolutional-neural-network object detection analyzes baggage X-ray images for prohibited or restricted items. The system presents item type and location to security staff for human review.
|
View source
|
| China | CHN Energy / Shuohuang Railway | Scheduling and Optimization |
Reported operational integrated shunting at Huanghua coal terminal combines AI vision, sensors, planning and communications to support locomotive movements and coal-unloading coordination. Pilot operation began in September 2022; the source reports formal operation from 22 March 2025.
|
View source
|
| China | Beijing Jiaotong University / Beijing Metro / Guangzhou Metro | Energy Efficiency |
Reported trial applications: an integrated energy-saving train-operation project includes AI driving knowledge and deep-reinforcement-learning adjustment algorithms. Beijing Jiaotong University reports trials of the overall technology on 17 metro lines in Beijing and Guangzhou. The source does not specify the AI component coverage on each line.
|
View source
|
| China | Shanghai Mission Information Technology (Mingxiao) / Nanjing metro stations | Customer Experience |
Reported station application: DeepSeek-based passenger information displays use retrieval-augmented generation and fine-tuning to answer spoken questions about transfers and local destinations in Nanjing metro stations. The supplier describes the AI dialogue function alongside navigation services.
|
View source
|