There may be an interesting difference in risk on this. AWe know in any given election some amount of votes cast are technically invalid for whatever reason. 

Switching out the current paper based system for some kind of software based system may be trading out known problems with individual invalid votes but lower degree of single point of failure systematic abuse for higher degree of certainty that any individual vote is correct, but higher chance of a long tail low probability but high impact abuse event

Transitioning to a software-based voting system may reduce individual errors but increase the risk of rare, high-impact abuses. Striking a balance between accuracy and security is crucial in electoral technology advancements.