The Chancellor Philip Hammond’s announcement that he wants to see fully driverless cars “on the roads” by 2021 has grabbed headlines. But in fact, depending on your definition of a road, we’re already there.

Fully driverless cars (ie. with no supervising driver) are already running in the UK. Trials were announced in four cities more than two years ago. The vehicles are effectively taxis, covering short, regular routes In low-traffic areas (the Milton Keynes cars were set to use the pavement, for example).

Perhaps what Hammond is talking about is fully driverless cars on “proper roads” – you know, the ones where almost 2,000 people die every year.

Again, a motorway trial has already been announced for 2019. But when it happens there’ll almost certainly be a driver on hand to take the wheel if things go wrong.

If the Chancellor really wants a fully driverless road-going future, the innovation won’t come from the hi-tech factories and the shiny labs – it will come from the mahogany-clad world of the lawyers, the regulators and the insurers’ boardrooms.

The question is simple: when there’s an accident, who takes the blame? (This question still stands even if driverless cars turn out to be safer than human-driven ones. After all, there’s nothing to stop a human driver crashing into a driverless vehicle).

If I’m no longer the “driver” of my vehicle, why should I take responsibility when something goes wrong? Instead, I’d want to pass the buck to my insurer. And you can bet the insurer’s going to be very keen to blame the car manufacturer.

(There are some interesting edge cases around responsibility here: for example, if I let the tyres deflate too much, will the car prevent me from driving, on the basis that the car doesn’t want to be liable for driving under my poor maintenance?)

I’m sure the technologists, given enough cash, can hit Hammond’s three-year target. I’ve got less faith in a speedy resolution by the insurers, regulators, and manufacturers’ lawyers…