Thread by @mattblaze

While ballot marking devices introduce security risks and tradeoffs that we need to understand better, this is not the most pressing problem or most productive election security hill to chose to die on right now. Getting Risk Limiting Audits in every state is.

Why? Because even the most rigorously voter-verified paper ballots don't help if the process that uses them is compromised. Risk Limiting Audits can confirm that the election outcome is correct, but only if we actually conduct them.

I believe the best currently available voting technology is (a) precinct counted, voter-marked paper ballots coupled with (b) risk limiting audits. We're making enormous progress on (a). (b) has gotten very little attention.

There are some terrible ballot marking devices on the market, as we discussed in the Voting Village report. We shouldn't use them, no one should buy them. But that's a tiny sideshow compared with RLAs in terms of overall election integrity.

This is why I find the activist focus on ballot marking devices so frustrating. While there are real issues there (especially with some current products), they're a sideshow compared with the larger challenges we need to be focusing on. We need rigorous post election audits.

There are also tradeoffs with BMDs that aren't well understood. While current BMDs appear to be pretty bad, it's not at all obvious that there couldn't be BMDs that actually help voters fill out complex ballots more accurately. The risks vs. benefits here are not yet clear.

This is probably too subtle to do justice to on Twitter. Some background:

My 2017 Congressional testimony: mattblaze.org/papers/blaze-g…

Our just-released report from the DEFCON Voting Village: media.defcon.org/DEF%20CON%2027…

The must-read National Academies study: nap.edu/catalog/25120/…

Getting Risk Limited Audits to become a standard practice is going to require an enormous amount of on-the-ground work and partnership between technologists at election officials at every level of government. The best way to build those relationships? Become a poll worker!

I should add that this thread and these links are highly US-centric. Some of the issues are universal to elections generally, but others arise from specific peculiarities of US elections.

This is apparently the thread that set off my current round of Internet swarming. I’m sorry if anyone feels attacked by something in my choice of words, but I stand by them. Risk limiting audits are utterly critical for assuring election integrity. They must be prioritized.

The underlying problem is that EVERY current (and likely future) paper ballot scanner can be easily compromised to alter the results. RLAs are the only known effective defense against this. Hand marked paper ballots don’t change this one bit. You need RLAs, too.