One of the paranoid, or perhaps not so paranoid, conspiracy theories doing the rounds is that the US Republicans manipulated election results through the use of Diebold voting machines.

I say “not so paranoid” because while I don’t accuse Diebold of committing any fraud, one of the problems with such voting machines is that they lack adequate auditing procedures.

Now computer scientists at RMIT have invented an encryption system that seems to do away with the concerns about auditing. I say “seems” because my head was spinning trying to understand what they had actually done. It is probably an explanation best experienced by seeing it in action rather than reading about it.

The beauty of it seems to be that it works on probability. You can never actually link up how a particular person voted (and that includes the person who voted), but if 2 percent of people confirm their codes, then it is enough to ensure that vote tampering couldn’t have happened.

Computerised voting is something that happens in the real world mainly on computer terminals at polling booths. One of the tantalising possibilities of the Internet is that you could offer voting at home. But one of the obstacles to that is the problem of insuring the integrity of the system. Perhaps a variation on this encryption system might do part of the job.