On 28 April, the Science Museum’s Information Age Gallery hosted the launch of GBRX’s Artificial Intelligence (AI) Action Plan. This was opened by GBRX’s Managing Director Toufic Machnouk whose key message was that AI will not simply happen in the rail system as this requires the development of strategic capability and the creation of new pathways for adoption.
Toufic welcomed Sir Tim Lawrence, chair of the Science Museum Group, who stressed that the museum’s mission is to preserve the past while helping society understand the future. As an example, he referred to the Marconi 2LO transmitter beside him which carried the BBC’s first daily radio broadcasts in 1922, bringing the information age into British homes.
Turning to AI, he recalled hosting the UK premiere of ‘The Thinking Game’, a film about a chess prodigy who used AI to model the human mind to understand diseases like dementia, shows how AI can consider the deepest questions of human health, cognition, and identity.
Next to speak was Rail Minister Lord Peter Hendy, who was unable to attend in person and so spoke live from Parliament which, he quipped, was another museum. He noted that the rail system’s longevity of assets, processes, and careers makes change difficult. He advised that GBRX’s AI Action Plan builds on the Government’s Transport AI Action Plan and that he was pleased to see real AI projects happening, while urging the industry to move fast.
AI Incubator
GBRX has established the AI Incubator Accelerator (AIIA) to accelerate the adoption of AI by helping the sharing of guidance on what works and building the capabilities that the industry needs to deploy AI at scale. AIIA will do this by:
- Working with partners to develop solutions.
- Providing targeted technical expertise for AI projects led by others.
- Providing advice including learnings from elsewhere in the rail sector.
- Sharing effective practice and supporting skills development.
Dr. Sarah Schlobohm, GBRX’s director of AIIA, explained that AI is not one thing but a messy ecosystem of techniques such as machine learning, optimisation, robotics, generative models, and classical algorithms. AIIA supports the implementation of AI pathfinder projects to overcome real world barriers and reach adoption and is to lead the following pathfinder projects. These take account of delivery and readiness across the rail industry and were showcased at the event by teams from across the industry:
- Passenger and customer experience.
- Network operations.
- Network planning.
- Rolling stock asset management.
- Infrastructure asset management.
- Organisational processes.
- Simplified ticketing.
In respect of the first pathfinder project, Dr. Schlobohm highlighted the pressing challenge of the UK’s labyrinthine fares system which has more than 400 ticket types and showed how ‘k means clustering’ can group similar fares to help simplify the system. She also gave a demonstration of a prototype Passenger AI Assistant which has been produced with input from the Rail Delivery Group which can plan journeys in natural language, identify railcard eligibility, suggest cheapest travel days, and provide real time station information.
Toufic closed the presentations by recalling that Tim Berners Lee’s original proposal for the World Wide Web was described as “vague but exciting”. He said it was also not clear how AI will shape our world. Yet it was clear that the rail industry must grasp the opportunities it presents and that this won’t happen by accident.
He noted that everyone present at this Science Museum event was surrounded by the work of people who shaped earlier periods of change. He then stressed that similar persistent leadership from within the industry will be needed to take full advantage of what AI has to offer.
Artificial Intelligence in Rail: The Industry Action Plan is available to view by following the QR code.
Image credit: David Shirres