Some of these ebikes are like tiny motorcycles ... that unfortunately riders are happy to ride all over the place.
I suppose stopping-distance might also be relevant, but there are more factors in there that are harder to measure.
a slow moving cargo bike has a lot of potential kinetic energy but it's moving slowly enough that a normal human reaction time can deal with it well and maneuver it safely. a bystander can notice the large bike that is moving slowly and make efforts to avoid it.
a much lighter bike going a much faster speed to equal the kinetic energy is a bigger threat to pedestrians up until a certain point.
in other words : a person on a bike doing 60mph is probably more likely to be acting recklessly than someone going a sane speed; thus it's an okay-ish proxy as a metric for a 'responsible driving' score.
Not as much as people tend to think, our intuition usually underestimates the importance of that squared velocity component. To illustrate, here are some vaguely-plausible numbers from some quick web-searches:
* A 65kg (~143 lb) person rides a big-bucket cargo bike of 65kg with 100kg of cargo, or 230kg total, moving at 5m/s (~11 mph) for 2875 joules kinetic energy.
* The same person on a 25kg passenger bike traveling at 10m/s (~22mph) is 4500 joules.
To be equal, the second rider would have to slow down to about 8m/s.
> a normal human reaction time can deal with it
True, it's easier for an alerted person to leap out of the way of a slower vehicle... but I think if everyone involved was perfectly alert and poised for action, we probably wouldn't be having this conversation.
Plus if you limited bikes to something reasonable like 20mph you'd have to start limiting cars to ~5mph.
That said, yes, there are some dangerous trends in car (un)safety to pedestrians.