Mile for mile, biking is significantly more dangerous than taking a car (as is walking), though not as dangerous as riding a motorcycle. (See here, or here if you're willing to tolerate older data and worse formatting.) You're trading off for physical fitness, though, and the life expectancy gained from increased fitness -- especially if expressed in terms of QALYs -- generally outweighs that lost from increased accident risk.

This undoubtedly varies by location but I haven't been able to find any good data on how. The waters are also muddied a bit by the fact that bike advocacy organizations like to promulgate the life expectancy data without the accident data -- not that I can really blame them, statistical literacy being what it is.