It’s probably also better if potential jurors don’t have the thought, “Well, I have some doubt but if this guy dies and I’m wrong, at least it’ll do a lot of good.”
1. If they have no issue with killing people.
2. have killed people
3. will continue to kill people
If all or even most of those victims being people who do not harm others then I'm alright with the murderers being used in experiments or killed themselves.
I don't care about them being found guilty or not, this is hypothetical and there are a number of ways that they could be found guilty when they haven't done it or not guilty when they have. Getting to the point where you can know without a single doubt that they in fact are serial killers is a different matter.
(1) is unknowable. (2) and (3) describe many, many people that are not generally considered to be serial killers (e.g., a soldier). If a 'known serial killer' were experimented on and died, (2) and (3) could also describe the researchers doing the experimentation.
We could modify (1) to say, "...and we have 100% certainty of this." But I guess really not certain what value the question has if we basically require 100% certain knowledge of the events and the internal state of the perpetrator. A 'no' answer would mean 'no' in the real world, too, but a 'yes' answer doesn't really tell us anything about the real world.
Somebody who is doubtlessly (and perhaps unrealistically) guilty of committing one of our highest crimes, for the sake of simplifying the question.