The issue with self-driving cars and public transportation is tricky. Self-driving cars might both drive down the cost of taxi-services, and drive up the convenience, both by not needing drivers and allowing providers to dynamically move cars to ideal position based on local demand, and so they might well drive up the use of personal transportation.
On the other end of the scale, while this would be possible with human drivers too, driving down the cost and response times of being able to automatically reposition and provision a fleet of cars would potentially make deep integration with public transport in ways that'd make public transport more attractive possible as well. E.g. imagine a public transport provider guaranteeing a maximum X minutes wait to drive demand, and plugging holes by dynamically provisioning cars to meet that guarantee on lower demand routes, or in case of full buses. Or an "end to end" service where you input where you want to go and instead of just getting the route, you're told to get on bus Y, and when you get of at station Z a car is there waiting to take you to your final destination.
Most of my Uber use is for journeys where public transport would be feasible but where being able to guarantee low waits and/or predictable arrival times makes a difference, and sometimes where a short leg (distance wise) of the public-transport option is too time consuming, and where it'd be "easy" to cut down on the car use if the transport system was better integrated.
If you can make the experience smooth enough to reduce the "end to end" convenience advantage of a car, even if it means using cars for part of it, it might well be a net improvement in public transport use.
On the other end of the scale, while this would be possible with human drivers too, driving down the cost and response times of being able to automatically reposition and provision a fleet of cars would potentially make deep integration with public transport in ways that'd make public transport more attractive possible as well. E.g. imagine a public transport provider guaranteeing a maximum X minutes wait to drive demand, and plugging holes by dynamically provisioning cars to meet that guarantee on lower demand routes, or in case of full buses. Or an "end to end" service where you input where you want to go and instead of just getting the route, you're told to get on bus Y, and when you get of at station Z a car is there waiting to take you to your final destination.
Most of my Uber use is for journeys where public transport would be feasible but where being able to guarantee low waits and/or predictable arrival times makes a difference, and sometimes where a short leg (distance wise) of the public-transport option is too time consuming, and where it'd be "easy" to cut down on the car use if the transport system was better integrated.
If you can make the experience smooth enough to reduce the "end to end" convenience advantage of a car, even if it means using cars for part of it, it might well be a net improvement in public transport use.