Earlier quoted context omitted.
Since zero crime is an impossible goal, the discussion is ultimately a question of how much assault by taxi drivers we should accept.
And we also have to consider how much we as a society lose by not having a taxi system with x amount of danger. 1000 people gaining a little bit quickly outweighs 1 person losing a lot (at least from a purely utilitarian perspective)
In my experience that assumption is usually false. There are usually ways where you can actually reduce the amount of danger and the number of people exposed to it and increase the number of people getting a more efficient service at the same time.
Look for these outward movements of the efficient frontier. The opportunity is almost always there. When we're really at the efficient frontier of what is possible and have no choice but to make tradeoffs let's say so loudly because something has gone very, very right to get to that point.