People working on fully autonomous vehicles tend to underplay the potential for imminent disruption, offering frequent assurances that drivers will be needed for decades to come.

Most analysts appear to believe the most significant obstacles will be regulatory and social. On the regulatory side, a transition is already underway, at least for semi-autonomous vehicles, and I don't believe regulation permitting fully-autonomous vehicles will be very far behind, at least not in the EU or the UK. In the US, with its fragmented legal system--and with trucking being the most important job in many states--it will likely take longer time. Wrt social obstacles, given the amount of protest that accompanied the rise of Uber, I think it's reasonable to expect a great deal of intense protests when the transition to fully-autonomous trucks begins. Trucking isn't unionized to the same extent many other sectors are, but I think there may be a push toward unionization in the next few years.

The intervening period will, I believe, be relatively calm. During that time, semi-autonomous trucks will make life considerably more pleasant for drivers, both employed and independent ones. They'll be able to relax more, get more rest, do paperwork etc. Owners of electric trucks will make large savings on fuel. Of course, companies will likely respond to the arrival of electric trucks by cutting compensation to independent drivers.

Given the successful early tests with long-haul platooning involving large caravans of semi-autonomous trucks, I think we'll see a rapid transition to driverless trucks under those circumstances, so long as the lead truck has a driver. Fully automated trucks are already in use outside of public roads, eg. mines and the like (see Rio Tinto for example).

There are a number of technological problems left to solve, eg. dealing with a large variety of types of cargo and loading configurations, varied driving environments (going from highway to city driving for example, which may be more difficult with a huge heavily loaded truck than it is with cars) etc. For large logistics companies, I think the issue of loading and unloading will be quickly solved, considering how far companies like Amazon have already come with their experiments in warehouse automation.

Ethical problems will be more difficult to address. When it comes to liability, a human driver may get a pass where a robot or its owner will be held to the highest standards with no room for error. A human driver may make a split-second decision to solve an immediate ethical dilemma (if it even registers as such) and humans will recognize the difficulty of acting appropriately in such a situation, whereas a robot making the same decision will be condemned.

At the same time, I don't think the cost aspect of this liability problem is likely to be a deterrent, because of the immense savings from going driverless. I think conservative estimates of savings to industry are at around $150-200bn. On top of that, you have savings to society at large. In 2014, close to four thousand people died in trucking-related accidents in the US, with another 111 thousand injured. In the vast majority of these cases, driver error seems to have been either directly responsible or a significant contributor to the accident. Even if you only automate the trucks, these numbers may be reduced to the point where liability issues would not be a major deterrent.

The savings and productivity gains are so immense that I think there will be a much greater impetus for changing regulations etc. The comparison to train-drivers, I think, is misguided, because of the magnitude of the expected savings.

I think the social aspects should be given some serious thought. With millions of poorly-educated full-time professional drivers at risk of losing their livelihoods, we should expect social repercussions. Many of these will find it difficult to find new jobs, and those who do may see substantial losses in income. I would expect another wave of populism, with truckers becoming the new and more influential successors of the coal-miners and oil-rig workers. Higher proportion of trucking jobs correlated with a larger margin of victory for Trump, and unless the economies of those trucking-dependent states change very soon, I'd expect similar outcomes in future elections.

Politics aside, there are other important social concerns, such as the prospect of hundreds of thousands of people ending up in poverty, or the high likelihood of increased suicide rates.