I think there are only two safe options if your intention is to avoid being assassinated: 1) don't tell anyone, 2) publish it anonymously. If you want to be able to prove to somebody that you're the author, sign the paper and keep the private key offline on a piece of paper.
It'd be interesting if it could be used to manipulate voting results, but e-voting is still in its infancy.
Interestingly, the consequence of releasing that algorithm would make having that private key insignificant since most schemas rely on factoring primes to secure the key pair.
It'd be interesting if it could be used to manipulate voting results, but e-voting is still in its infancy.