I've had two automobile-involved crash in about 8 years of heavy bike commuting. Both were right hooks -- one I flipped my bike and sprained both elbows avoiding the car, one I ran into the car. Both of them were within the first year, and I have a pretty good idea how I could have better anticipated the accident (and probably avoided it) by paying more attention to the situation around me. That doesn't mean I think I'll never have another crash involving an automobile, but I do think I'm safer now than when I was a newbie.

As for whether or not it's inevitable, I like this link:

http://cyclehelmets.org/1026.html

It's well known that you can lie with statistics. One of my favorite statistical lies is to look at the second chart on that link and claim that I'm six times safer when I'm on the bike ["Bicycling" = 0.26 fatalities per million hours] than when I'm not on the bike ["Living (all causes of death)" = 1.53 fatalities per million hours]. That's a lie, of course, but it does put the relative risk in some perspective. It also gives you a good excuse to introduce the health benefits of cycling into the safety discussion.

I also like that cycling, while apparently not quite as safe as golf, is pretty close and is not nearly as dangerous as tennis.