MIT Security Analysis of “Voatz” App [pdf]
internetpolicy.mit.edu
internetpolicy.mit.edu
I think the goal should be that there's not a net increase in risk of election tampering by moving from vote technology A to vote technology B.
I am inclined to think that the peak might be filling out paper ballots to be read by machines (penciling in ovals/scantron).
If we want to improve turnout, we should try some existing ideas that are much less novel and risky: offer early voting, make elections a holiday or hold them on a weekend. Or make it compulsory.