There are many levels voting can be corrupted (for example, UI could be showing to users some voting options, while server would be storing something else, because client code served by the client would be compromised). But one simpler way to make at partial accountability is for server to issue a proof of results. So server could have a key pair and then sign a proof with its private key so that anyone can verify it with the public key. The proof could contain all the input to the voting algorithm and the results. One could maybe download such proof at any point for every motion.
The idea is that if users store such proofs if at any later point somebody changes results, they will mismatch the stored proof. And users could not fake those proofs, so server had to issue them. The downside is that in our current system, attacker could simply change votes and claim that users changed their votes at the last moment, and not change previously stored votes. So it is unclear how useful would such proofs really be.
There are many levels voting can be corrupted (for example, UI could be showing to users some voting options, while server would be storing something else, because client code served by the client would be compromised). But one simpler way to make at partial accountability is for server to issue a proof of results. So server could have a key pair and then sign a proof with its private key so that anyone can verify it with the public key. The proof could contain all the input to the voting algorithm and the results. One could maybe download such proof at any point for every motion.
The idea is that if users store such proofs if at any later point somebody changes results, they will mismatch the stored proof. And users could not fake those proofs, so server had to issue them. The downside is that in our current system, attacker could simply change votes and claim that users changed their votes at the last moment, and not change previously stored votes. So it is unclear how useful would such proofs really be.