in the US, this would be insurance that the scientist (and/or their emloyer) pays for. if enough scientists perform risky experiments that it meaningfully increases payouts, I'm sure the insurer could figure out a way to account for that in the premiums.

in short: who cares?