Ok, so I was looking for motorcycle ownership and moped ownership rates and didn't make much headwind with google so I estimated ownership by counting how many motorcycles and how many mopeds are being sold on craigslist.
It appears 1 moped is offered for every 20 motorcycles, so I would think motorcycles outnumber mopeds 20:1.
So it would appear that motorcycles have about 5 times the fatality rate than mopeds per year. However, motorcycles on craigslist seem to have more far more mileage than mopeds (10,000 average compared to 1000 average) suggesting motorcycles travel much further so, per mile, motorcycles might actually be twice as safe.
That sounds like a bogus inference, though, so I'll scrap it. I would think that the fatality rate would be either be parabolic or non-decreasing, meaning the moped fatality rate would be less than the bicycle and motorcycle fatality rate, or greater than the bicycle fatality rate but less than the motorcycle fatality rate. Since it appears motorcycles account for 5-8% of fatalities while accounting for 2%, it seems like mopeds would have a fatality rate less than that. So, at max, ones "average" lifetime chance of death from moped would be 1/40 at max and 1/320 at a minimum. Ok, so that doesn't help much.
Oh wait... using the 5-8% motorcycle fatality figure, and assuming that motorcycles have about 5 times the fatality rate of mopeds, it appears that mopeds have a fatality rate similar to cars. BUT, people don't use their mopeds nearly as much as people use cars judging from mileage, so the per mile fatality rate seems like it might be 5 times as much as cars (Since motorcycles have a per mile fatality rate ~25 times much as cars, and the fatality rate of mopeds seems to be 5 times less). Since the bicycling fatality rate per mile is nearly 2.5 times that of cars, it seems like mopeds would be nearly twice as dangerous per mile than bicycles. What a concept - more speed, more danger.
Now, if one were to assume "ebikes" were similar to mopeds in terms of performance, then it seems they have nearly twice the fatality rate per mile than bicycles. But supposing one travels less than (10,000-miles-per-average-car/5-times-less-safe) ~= 2000 miles per year on a moped-like ebike, it seems one might be fatality-based safer than a car in a lifetime. Those injuries can suck, though, such as dogman's. However, I believe typical car injuries are no cakewalk either.
That assessment entirely depends on the motorcycle/moped use rate, which I might not have accurately inferred.
Edit: I don't believe my sources are good enough, so I think this post probably contains a lot of garbage.