AI and rail fares knowledge

[D1537]

Moderator note: split from

"Edge case" concerning Oyster 60+ relating to a Thameslink 9:28 train

I've had one of these for 5 years now. I'm aware there are several questions about "touching in" with Oyster 60+ but haven't found an answer here to my specific question. I live very close to a South London Zone 4 Thameslink station, Ravensbourne. Occasionally when I go into (central) London...

www.railforums.co.uk

It's amusing that AIs can do astonishing things in areas such as research and scientific development but can't negotiate their way through the complexities of UK rail fares.

This probably says more about the nonsensically complicated rail fare structure than it does about AIs.

 

[Haywain]

This probably says more about the nonsensically complicated rail fare structure than it does about AIs.

There are different versions of AI. Scientific research is rather less likely to be using a free version. There's an old adage about getting what you pay for.

 

[yorkie]

It's amusing that AIs can do astonishing things in areas such as research and scientific development but can't negotiate their way through the complexities of UK rail fares.

I'm not sure that AI can't be used for this; has anyone specifically trained an AI model for this purpose in the way that has been done for scientific development?

This probably says more about the nonsensically complicated rail fare structure than it does about AIs.

Simplification in rail fares usually means tightening up restrictions, e.g. it could be simplified by not allowing any passes, railcards or off peak fares before 0930.

It wouldn't be simplified the other way round; and indeed it couldn't be done without increasing subsidy. I would be happy to pay increased taxes to pay for it, but I don't think it would be considered a vote winner as the UK is too car-centric sadly

 

[Crithylum]

I’ve had a thought that making a journey planner / fare finder would be a very good AI benchmark. Only providing the electronic data and the document explaining the protocols would leave a lot for the AI to work out.

 

[yorkie]

I’ve had a thought that making a journey planner / fare finder would be a very good AI benchmark. Only providing the electronic data and the document explaining the protocols would leave a lot for the AI to work out.

Good luck to anyone who does this, with the various parts that aren't in the electronic data, such as Elizabeth Line validity, Avanti Standard Premium, LNER 70min flex, and many, many more

 

[Trackman]

Also, asking AI about route planning is a right laugh!