A new action plan setting out a sector-specific approach to adopting artificial intelligence (AI) in Britain’s complex and safety-critical railway system has been launched by GBRX, the strategic innovation and technology arm of the future Great British Railways (GBR).


Artificial Intelligence in Rail: The Industry Action Plan was launched at an event held at the Science Museum in London on April 28. The plan focuses on overcoming long-standing barriers to the widespread adoption of AI across Britain’s rail sector, such as incomplete or inconsistent data, governance and assurance frameworks, and a lack of AI skills and organisational readiness.


Development of the action plan has been led by GBRX, and has involved collaboration across the sector, including rolling stock manufacturers, open-access and freight operators, as well as the passenger operators now returning to the public sector, which along with infrastructure manager Network Rail will form GBR.


Key to moving beyond isolated AI pilots to a coordinated, sector-wide approach will be an initial set of priority pathfinder projects in the following areas:


  • passenger and customer experience
  • network operations
  • network planning
  • rolling stock asset management
  • infrastructure asset management, and
  • organisational processes.

Initial projects will be refined and prioritised through a process of progressive review, led by the Artificial Intelligence Incubator Accelerator (AIIA) established by GBRX, which says it is the first of its kind in the British rail sector.


“Across the sector we are already seeing pockets of excellence in how AI is supporting operations, engineering and customer experience,” says AIIA director, Dr Sarah Schlobohm.


“AIIA exists to accelerate that progress by helping teams move through pilot, production and operational use. We share guidance on what works, and build the capabilities the industry needs to deploy AI at scale.”


In the field of passenger and customer experience, Schlobohm highlights the potential for machine learning to assist with the process of fares, ticketing and retail reform that is now underway in preparation for the creation of GBR. In place of operator-specific websites and alongside third-party retailers, GBR will offer passengers a single unified online point of contact for journey planning, ticket sales and passenger assistance, based on official data feeds.


Over 400 different fares and over 55 million possible ticket combinations are currently available, according Schlobohm, with work now underway to create a new integrated fares model that is a priority pathfinder project for AI deployment. The new model aims to make fares fairer and easier to understand by removing confusing choices, with a maximum of three fare types for every journey.


“AI presents a significant opportunity to improve how the railway works for people,” says GBRX managing director, Toufic Machnouk.


“In a system as complex and safety-critical as ours, it requires deliberate action to create the conditions for adoption. This plan brings the sectors together to do exactly that, creating the capabilities and pathways that advance the role of AI in how the railway is planned, operated and experienced, delivering improvements for passengers and freight.”