Self-driving vehicles are expected to carry passengers on public roads within the next decade.

In the near future, cars without a driver will probably carry passengers on normal roads. In my opinion, this will be a positive development, although it will also create some problems.

The main advantage is safety. Most accidents happen because of human mistakes: the driver is tired, he looks at his telephone, he drives too fast or he drinks alcohol. A computer never becomes tired and never loses concentration, so the number of accidents and deaths on the roads should become much smaller.

Another benefit is for people who cannot drive. Old people, blind people and people with a disability often cannot leave their home alone, and they depend on their family. With an automatic car, they can go to the hospital or to the shop whenever they want, and this gives them independence.

However, there are also disadvantages. Millions of people work as taxi drivers, bus drivers or lorry drivers, and all of them can lose their job in a few years. It is very difficult for a fifty-year-old lorry driver to find a completely new profession.

In addition, there is a legal question. If an automatic car has an accident, who is guilty: the owner, the company that made the car, or the programmer? Governments must answer this question before these cars appear on the roads.

In conclusion, I believe self-driving vehicles will be positive because they will save many lives and help people who cannot drive, but governments must prepare new laws and help the drivers who will lose their work.

Vehicles capable of driving themselves are expected on public roads within a decade. I regard this as a positive development overall, though the obstacles to it are less technical than they are legal and social.

The case rests principally on safety, and it is stronger than it first appears. The overwhelming majority of collisions are caused by failures of attention — distraction, fatigue, intoxication, misjudged speed — and these are precisely the failures a machine does not have. A system does not become bored on a motorway at three in the morning or check a message at a junction. Given that road deaths run into the hundreds of thousands annually worldwide, even a partial reduction represents a very large number of lives. There is a second benefit that receives less attention: people who cannot drive because of age, sight or disability would gain independent mobility they currently lack entirely.

The difficulties are real but different in kind. Driving is among the most common occupations for men in many countries, so the displacement would be unusually concentrated — not spread thinly across an economy but removing one job from one group, often in middle age and in regions with few alternatives. Liability is equally unsettled: when no one was driving, responsibility for a collision has to fall on a manufacturer or a software provider, and no legal system has fully worked out how.

In conclusion, I believe the development is positive, because a technology addressing the exact cause of most road deaths is difficult to argue against. Whether the benefit is realised depends on resolving liability and planning for the workers affected rather than on the engineering.

The strongest argument for autonomous vehicles is not that they are impressive but that they attack the specific cause of most road deaths. Roughly nine in ten collisions originate in a failure of human attention — fatigue, distraction, intoxication, misjudgement — and these are the failure modes that do not apply to a machine. A system does not tire on a night shift, does not glance at a message, and does not overestimate its own reaction time. I therefore regard the development as positive, though I think it will be delayed by something other than engineering.

The scale involved makes the case unusually strong. Road traffic kills well over a million people a year worldwide and injures many times that number; it is among the leading causes of death for young adults. A technology that reduced this even by half would rank among the significant public health achievements of the century. Alongside that sits a benefit rarely counted: people excluded from driving by age, blindness or disability would acquire independent mobility, which for many is the difference between an autonomous life and a dependent one.

The costs are genuine and unusually concentrated. Driving is one of the largest single occupations for men without university education in many countries, and unlike most technological displacement — which spreads thinly across sectors over decades — this would remove one occupation, from one demographic, frequently in middle age, and often in regions where alternative work is scarce. Treating that as an unfortunate side effect of progress is precisely how previous transitions produced lasting damage.

What I expect to be decisive, however, is neither of these. It is that autonomous vehicles will be judged against a standard human drivers have never had to meet. Human road deaths are statistical and diffuse; they occur constantly and are almost never news. A single death caused by a machine is salient, attributable and intensely reported, and no manufacturer has a defence that survives contact with a grieving family. The consequence is that a system demonstrably safer than the average driver may still be politically unacceptable, because the deaths it prevents are invisible and the ones it causes are not. Liability law compounds this: with nobody driving, responsibility must attach to a company with substantial assets, which raises the cost of every remaining accident considerably.

My verdict, then, is clearly positive, with a qualification about what determines the outcome. The engineering will arrive; whether the benefit does depends on whether societies can accept a smaller number of visible deaths in exchange for a much larger number of invisible ones — which is a question about how we weigh evidence, not about technology.