I Stand with Avi (regarding American Idol)

First, I need to express solidarity with Avi, who went out on a limb and professed his love of American Idol. I, too, am a closet American Idol fan, and I completely agree: if only the voting were verifiable! I personally think Jared should have qualified for the final 12.

But on to the immediate issues, regarding my post yesterday and Avi’s response. I understand that you can’t recommend open-audit machines for immediate adoption… although I believe that technology like VoteHere‘s should be considered at the very least for prototype deployments.

I wonder, though, how will we ever vote on such open-audit systems if people aren’t aware of them? Worse, how can industry and research ever be expected to improve on open-audit designs if overly prescriptive laws simply disallow them on the basis that they don’t produce a VVPAT? Would it be too confusing to say, at a hearing, “Cryptography can secure computer-based voting, which is why we really need to evaluate these systems, encourage research on the topic, and write laws that leave open that possibility. That said, if you need to buy a voting machine today, buy an opscan with disclosed source, because none of the current computer-based voting machines come even close to the auditability standard we need, as specified by NIST.”

What bugs me specifically is the oversimplification of NIST’s software independence concept. Rivest and Wack carefully crafted their paper to state the crux of the issue: we don’t want to trust any single piece of software with the keys to the kingdom. The beauty of this phrasing is that it lets you say “Hey, Big Company, does the correctness of the voting result depend on your software? Yes? That sucks!” As best as I can tell, the current Holt Bill oversimplifies the issue and mandates a paper trail. Why not, instead, mandate software independence, as defined by Rivest and Wack? Why not reference the expert work of NIST/ACCURATE/VTP in these bills?

UPDATE: I almost forgot that I wrote about my American Idol addiction and its relation to crypto last year.

4 thoughts on “I Stand with Avi (regarding American Idol)

  1. It would seem that American Idol is a great test bed for cryptographic voting, no? We wouldn’t be able to do much about DoS and other voting risks similar to internet voting, but we should be able to convince them, for a modest investment in their voting infrastructure, they could “prove” that America did vote a given singer out. I could imagine, if the network is in control of the whole thing anyway for marketing/advertising purposes, that this might not fly. But if I were at VoteHere, etc. I’d be seriously thinking about how to pitch this kind of concept to the AI peeps. We need more test beds for new voting methods besides union and student elections.

  2. It would seem that American Idol is a great test bed for cryptographic voting, no? We wouldn’t be able to do much about DoS and other voting risks similar to internet voting, but we should be able to convince them, for a modest investment in their voting infrastructure, they could “prove” that America did vote a given singer out. I could imagine, if the network is in control of the whole thing anyway for marketing/advertising purposes, that this might not fly. But if I were at VoteHere, etc. I’d be seriously thinking about how to pitch this kind of concept to the AI peeps. We need more test beds for new voting methods besides union and student elections.

  3. Joe, I do think it would be cool, especially since this is all doable easily via homomorphic tallies. That said, you can vote multiple times on American Idol, and you never authenticate… so I’m not sure how much verifiability you would really get from this…

  4. Joe, I do think it would be cool, especially since this is all doable easily via homomorphic tallies. That said, you can vote multiple times on American Idol, and you never authenticate… so I’m not sure how much verifiability you would really get from this…

Comments are closed.