people will pay for it. People who have to drive 100 miles on a motorway regularly will jump at the chance to sit back, inspect their eyelids and listen to smooth Fm instead of stressing about doing 50 in all the roadworks.
No one is suggesting that trains CANNOT be automated. Of course they can. They already are. It's the processes that are required to reach that goal that are the issue.
We have a large complex OLD railway system. To retro fit the whole lot will cost a lot of money and take a lot of time. Someone has to be willing to stump up the cash to do it. At present the state pays for most infrastructure work because the risk to private capital is too great. The trains are run by private operators. They will pay for the ongoing costs of staff because that's still cheaper than the sums needed to change the infrastructure the trains use are too great for them to obtain a return within the terms of the franchises.
So when we build NEW lines - HS1, the DLR, the Victoria Line - we can automate from the start. That's much easier. The vast majority of automated trains run on NEW segregated lines. Unless we have a major rebuild of the network, those are always going to be in the minority. You may also notice we seem to have a big problem with such projects in the UK - the West Coast upgrade and the Thameslink programme - ATO worked for a demo train and actually not very well anyway - and that was just on a short stretch.
Yes, we can do it incrementally. That's still a long term process and may never be completed - we have been trying to electrify our trunk routes since 1955 and we still haven't reached Bristol, Cardiff, Nottingham or Sheffield. Meanwhile you still have to employ drivers to cover the non automated bits in the same way that you still need diesel trains until all the wires are up. More hefty costs.
Meanwhile there is a review and a new policy on rail roughly every two minutes, so as soon as one project gets underway (London Overground say) it gets stalled by a political change, or the financial climate changes and projects get slashed.
Neither trains nor driverless cars can defeat the laws of physics. Please don't dismiss the leaf fall problem - you should be aware now that it is a real problem, and the railway takes the measures that are available to it to try and address the problem. Even when it goes for broke it is thwarted, our SoS recently stopped Network Rail cutting down trees close to a line because of NIMBYs complaining. You can't have MORE trains on a network that has been brutally rationalised and still maintain it properly and safely in the brief window overnight or at weekends. Something has to give. And driverless cars will still have to do 50 through miles of road works and sit in traffic jams.
Technology is already replacing people. There are platform indicators instead of a man putting up a wooden board - though some might argue the latter worked better in disruption. Ticket machines mean that booking clerks now leave the office and sweep platforms and clean toilets - though not all tickets are available and it is not always clear what the right ticket is. CCTV means most of London and the SE is DOO now - and I can assure you from experience both as a former commuter and a current DOO driver that it is definitely a retrograde step. Ironically DOO makes automation less likely in the fragmented system we have - a mess of different operators, traction types, no clear direction or central management, because until the whole thing is automated, you need to pay drivers anyway, even if you will eventually be paying them to be guards instead, so there's no real incentive to lose them.
So yes, we can automate the railway. IF there is a will to see it through. IF it is properly resourced. IF there is an effort to get the technology right. IF we up our game with project management. IF rail is taken seriously as part of our national infrastructure. It's interesting that in other European nations that do value their railways properly, with continuous programmes of investment and improvement, haven't seen any major implementation of driverless trains yet. The practical versus the possible.