IMHO, anything that not enough people can find interesting and engaging should be automated — injury risk is also a great reason to automate something. Of course, there is a cost involved, so we shouldn't expect any of it to happen overnight.
While your point is true that every profession has an injury risk, it's the type of comment that misses the point: there is a significant difference in those stats, and by trying to equate them, you are purposely watering down the significance of one where injuries are more prevalent (or at least harder).
It's a response similar in style to deflecting from attempts to solve one crisis by introducing another crisis.
My point, was that both physical/labor intensive jobs and office jobs bring their own types of workplace related problems. And just because the job may carry with it a risk of physical injury doesn't mean the job should be automated. Afterall, any type of movement the body does carries with it a risk of injury.