Driverless vehicles are undoubtedly coming, but how will they change our cities? There’re two eras here; the transition period when driverless and conventional vehicles share road space, and the potential utopia (or dystopia?) when all cars are fully autonomous.
The transition would be enormously complex, both because human drivers are imperfect and because the algorithms controlling driverless vehicles will have difficulty dealing with the sometimes capricious and fickle nature of human drivers and pedestrians.
I’m going to take the easier option and consider a world where driverless cars have completely replaced human-controlled cars. In this world, many of the problems that loom large at present – like legal liability, ethical choices, hacking, imperfect algorithms – have been solved.
The main promised benefits from a fully driverless world are:
The near elimination of crashes and the associated personal and material costs
Significantly lower travel costs, mostly from the ability to use in-vehicle time in productive and enjoyable ways
An increase in the traffic capacity of roads due to closer travelling
Parking facilities shifted to less strategically valuable locations
It’s also often argued that fewer vehicles would need to be manufactured and fewer parking spaces provided because driverless vehicles could be shared. However, that’s not down to driverless cars; we could already enjoy most of those benefits if we chose to e.g. via taxis and share cars. It’s more accurate to say driverless vehicles would provide the opportunity, or the pretext, to adopt a new model based on sharing.
Here are some of the key ways a fully driverless city might be different from today’s world:
We might travel more because time spent in-vehicle would be much less onerous if you could sleep, work, or do pretty much anything, in private. Moreover, there’d be a temptation to send the car off on errands without a passenger e.g. to fully automated shops. This would work against the promised increase in road capacity.
We might live further away from work and other destinations, again because the “cost” of travel would be lower. A potential implication is an expansion in regional living, although this might be counteracted to some extent by the growing preference for cosmopolitanism. The already flawed notion of the “20-minute city” would look even sillier.
High capacity public transport would still be necessary to shift large numbers of travellers e.g. to and from the CBD. Driverless trains should cost less to operate, permitting very high frequencies. Shared driverless vehicles operated by the transit manager should provide a much more attractive way of travelling between home and the station than conventional buses, as well as improving cross-town services.
Lower travel costs might encourage firms to locate in centres where they could benefit from agglomeration economies. In that event, many jobs in dispersed suburban locations should migrate to the CBD and a select number of larger suburban job centres.
Lower freight costs should reduce the size of trucks and vans required within cities.
There might be tighter regulation of activity and land uses along major roads so driverless vehicles have an easier environment to ‘read’. This is more likely to be a compromise dating from the transition period and should disappear as AI gets better.
There might be a large net loss in jobs, especially semi-skilled jobs like truck and bus drivers. At least some of those losses should be offset by job gains in other areas where money previously spent on drivers would now be spent instead.
The problem of additional travel could be addressed if the existing model of private ownership of cars were replaced by a sharing model akin to the way taxis currently operate:
Driverless vehicles could only be “rented” as needed by users; this would reduce the size of the fleet and the number of parking spaces required
Travel is paid for by the kilometre/hour and weighted by the level of congestion; this would moderate usage by making the cost apparent.