The myth of driverless cars


We may have a lot of unfilled potholes and cutbacks in funding for buses, but there always seems to be money available for research projects on autonomous vehicles.

This latest tranche of £22m of grants brings the total up to £120m for 73 schemes. Barely a day passes without predictions that these vehicles are the future and will soon be dominating the road environment.


Indeed, money for transport research and development is being almost entirely focused on connected and autonomous vehicles. One researcher, who did not dare to be named, told me: ‘You only have to mention the word “autonomous” in a research bid, and the money comes through.’

This is already distorting the market and is set to get worse as politicians jump on to this bandwagon, as illustrated by chancellor Philip Hammond stating last year that driverless cars will be on UK roads by 2021.

Yet, there has been no proper debate about the desirability or feasibility of these vehicles, let alone a thorough analysis of their technical capability. These vehicles are being developed by a combination of auto manufacturers and tech companies, like Google, whose subsidiary Waymo is the biggest player. As with many high tech inventions, there is no public demand for these products and consequently these companies are using the media to make outlandish claims about their desirability and imminent arrival, which do not stand up to scrutiny.

Initially, the companies producing these vehicles envisaged that they would be a natural extension of existing ones. This proved to be a mistake because testing showed that reducing drivers to mere occasional oversight meant that they were not sufficiently alert when they needed to intervene. Since one of the key selling points of this concept is increased safety, the developers have been forced to go straight to Level 4 capability (out of six levels from 0 to 5), at which point cars are able to drive themselves in all situations with no human intervention.

Technically, so far this has proved insuperable. Despite all the hype surrounding trials, these have been limited to relatively simple situations in specific geographic areas and good weather conditions.

Moreover, there has nearly always been an operator ready to take over in dangerous situations where the auto pilot has failed for instance and the occupants or other road users were put at risk.

In order to convince the public of the benefits of autonomy, the developers have suggested that the cars would be electric pods that would be shared rather than individually owned. They have put forward this model to highlight the considerable savings from the fact that people would not need to buy their vehicles, and that parking provision would no longer be necessary either at home or in workplaces. Waymo and Uber, in particular, have been pushing this model.

However, it is fundamentally flawed. While people in urban areas might have instant access to one of these pods, it would not be economically or practically feasible to provide them on demand in rural or sparsely populated suburban areas. Moreover, people have specific types of cars to accommodate their requirements, whether they be keeping their golf clubs in the boot, having seats for children or space for tools, or simply having a small car in order to reduce fuel consumption. The idea that standardised pods fulfil all requirements is fanciful.

Then there are issues about the capability of these vehicles: how would they distinguish between a traffic jam and a row of parked cars? What would happen when two pods met each other on a single carriageway road and the occupants were not able to reverse them? Would legislation against ‘jaywalking’ be necessary as the vehicles would have to stop if someone with evil intent – or simply in a hurry – stepped in front of the vehicle? And as for Uber, which currently relies on owner-drivers, can the company really afford to buy all these driverless cars since there will be no drivers to own them? I asked an Uber executive this question recently, and he responded that he expected the car manufacturers to provide the vehicles for free on the basis that they would then earn an income from the taxi trips. I responded that this seemed an unlikely scenario, to say the least.


The list of practical questions is almost endless, and so are issues relating to the regulatory, legal, insurance and software security considerations. The more that one analyses this phenomenon, the less realistic it appears.

The whole concept seems to be borne of the needs of the tech companies to find some use for their monopoly profits with the support of the auto manufacturers who are terrified of being left behind by their rivals.

Unfortunately, as ever with the tech companies, they present this development as benign – it will improve safety and relieve people of the burden of driving – when, in fact, the only motive seems to be creating a product to ensure their continued profitability. After all, self-driving cars will allow people to spend more time using Google products.

One could argue that a few million pounds of government money wasted on gadgetry is trivial but, in fact, there are numerous damaging effects. Firstly, researchers in other fields of transport, such as improving information systems or making buses more fuel efficient are aghast that the limited funds available for government support of research and development are being wasted on these boys’ toys.

Secondly, the hype which these grants help to stimulate encourages the view that autonomous vehicles will soon appear on the roads and therefore allows their supporters to argue that spending on alternatives, such as improved public transport, is a waste. 

Thirdly, these grants appear to go against the grain of the free market ideology so favoured by the current government. If this technology genuinely represents a possible future for the transport system, then it is up to the free market to stimulate it. Clearly, however, there is no business model and it is unlikely one will emerge. None of the cars so far on trial have been priced as they are not yet on the market but it unlikely that any would cost less than a six figure sum. Even mass production might not make them affordable. The sensors and other equipment they require are incredibly expensive.

The immensity of the task of creating vehicles capable of self-driving in all weathers, on all types of roads (and even down the off-road lane to the golf club), and in situations with large numbers of pedestrians may mean that these vehicles will never be able to be used in the way their protagonists suggest. Some tech, indeed, seems to be going backwards. A friend who just bought a Range Rover says that its ‘self-parking’ facility previously fitted as standard has been removed from new models because it never did work satisfactorily.

True driverless cars may, therefore, not be feasible and therefore the advantages, such as freeing up central city parking, reducing road casualties and allowing non-drivers such as the children, the infirm and the blind to have access to cars, will never be delivered. Politicians must take note and not be conned by the hype. They must not allow transport policy to be determined by wishful thinking by the tech companies and their allies in the automobile industry.

Signed copies of Christian Wolmar’s book, Driverless cars: on a road to nowhere are available for £10 post free from the author or just go to Amazon.


Also see

Register now for full access

Register just once to get unrestricted, real-time coverage of the issues and challenges facing UK transport and highways engineers.

Full website content includes the latest news, exclusive commentary from leading industry figures and detailed topical analysis of the highways, transportation, environment and place-shaping sectors. Use the link below to register your details for full, free access.

Already a registered? Login

comments powered by Disqus