Conundrum for the day

Here is a philosophical problem I am working on that goes well with my post on immortality. Your comments are always welcome:

In my lifetime as a driver, I stand some (fairly low) chance of killing an innocent pedestrian. Few people would argue that I should be prohibited from driving. Assume, however, that science prolongs (fit) human life forever, at least unless you are struck down by a car. My chance of killing an innocent pedestrian then would approach certainty, given that I plan to continue driving throughout an eternal life. In fact I could be expected to kill very many pedestrians. Should I then be prohibited from driving? When we make a prohibition decision, should we measure the risk of a single act of driving, or the risk of driving throughout a lifetime? Measuring the bundled risk appears to imply absurd consequences, such as banning driving for people with sufficiently long lives.

Alternatively, measuring the risk of only the single act is vulnerable to counterexamples. Imagine an involuntary game of Russian roulette with very many chambers in the gun, played very many times against me. The chance of my death from any single firing is very small, but surely we would prohibit such a game, looking at the high overall risk of the bundle. In this case we consider the bundled risk, but does this mean that we should stop immortals from driving cars?