It's a great joke, but a poor strategy. Remember the end game: minimize harm. The universal solutions that remove human error have a better long-run potential to minimize harm, since they can exceed the effectiveness of even the most attentive and cautious human driver.

There was a coach bus crash in the Alps a few years ago, and in the investigation they found out that one of the factors contributing to the crash was that the driver only had few years of experience, and only ever drove modern, top of the line buses. The problem with those is that they drive as easily as a small hatchback, despite being 20 meters long and weighing 17 tonnes. The engines are powerful, the transmissions are usually automatic, the brakes are strong - so you can easily forget that you are driving a 17-tonne fridge, and no matter how good your brakes are, you can't be braking continuously on a 15km long mountain descent, because the whole system will fail. The bus was equiped with engine and magnetic retarder for the drive, but like I said, on a modern bus that drives this easily, the driver just never had any need to use them. I think the report concluded that a driver with more experience in driving older vehicles, where you couldn't rely on assisted disc brakes, where all you had was a set of brake drums and just had to know how to use the 2 other braking systems, would have saved the bus.

Similarily, airplane pilots spend so much time in autopilot mode that they forget how to operate a plane in emergency situations.

10 accidents where everybody walks away is much better than 1 accident where someone dies.

20 years ago they redid the traffic pattern where I used to live to something many people found really confusing and many people claimed felt dangerous. A study 2 years later found that, while it was true that accidents on that part of the road had increased, the number of injuries due to accidents had plummeted. A follow up study a few years later found that accidents had dropped back to around their original level and serious injuries where still basically at zero.

The problem with this theory is that the rate of fatalities has gone sharply down as cars have become safer. There may well be some degree of compensation where people drive in a more dangerous manner, but the added safety far more than compensates for it.

Those same drivers are, unfortunately, also likely to be those least amenable to actually using it. Some people I know have expressed disgust toward the idea of autonomous driving; coincidentally, almost all are horrible drivers that I wouldn't want to be with as a passenger.

Other forms of transportation have better trained pilots. :)