Professor David Dill, Stanford professor of computer scientist recently was awarded a large grant from the National Science Foundation to study voting systems. He gave a presentation at the U of Utah Engineering Dept on Fri. Sep. 16, 2005. The room was packed primarily math/science students, professors and a few election activists.

Hi-lights of Dill's speech:

"Man Behind the Screen" Voting System:

DREs could easily switch votes on election day to cast them for a different candidate than voters expect. Yet, People who support DREs claim that they are perfect and never lose or change votes. Current technology is not up to the task.

Paper ballots are the only proven option for ensuring that our votes are cast and counted as intended.

DRE paper rolls should be audited with the same method that the voters verify them (i.e. by hand - not the barcode).

Facts:

Diebold DREs use wireless technology (as mentioned below in the U.S. CERT warning). This broadens the opportunities for tampering.

We cannot verify that the desired software is running on our voting machines, even with stringent software design/review.

We don't know how to eliminate program bugs. With inspection and testing, only the most obscure, unpredictable bugs remain. In one case, three separate "public totals" on the same printout gave three different counts on a tabulator: 947, 945, and 1019. It turned out that there had been three copies added of the votes from one machine, but even that one machine gave 2 different vote totals in its redundant memories.

The new risks of DREs to voting systems are that wholesale, nationwide fraud can affect votes nationwide and even honest local election officials have no way to stop such tampering.

Banking systems are much easier to secure than voting systems because we must discard vital information in voting (the connection between the voter and the ballot) and in banking (unlike voting):
  1. fraud is quantifiable
  2. fraud is discoverable
  3. fraud is measurable
  4. banks audit their paper records
  5. banks use paper backup & crime still occurs
  6. customers are protected

What assets are protected by honest accurate elections? The national budget is trillions of dollars. Even at a local county level control of a big county can mean control of land use decisions and contracts in the millions. Who are the potential attackers? Hackers, candidates, zeolots, foreign governments, criminal organizations (not surprising that many felons are in the employ of some voting machine vendor companies).

Wouldn't it work far better for the man behind the screen to cast our votes for us, then hand us the ballot so that we can verify it, and then carry it ourselves to a ballot box so that we know it was correctly cast. (more like the AutoMARK does).

David Dill favors the AutoMARK and optical scan ballot systems for counting votes like we recommended to Utah.

The problem with unaudited e-voting is that we can't prove that the vote counts are correct. So why should we trust these machines?

Someone asked Dill why election officials are supporting DRE voting machines - laziness or dishonesty? David said that it is dangerous to speculate about motivations, but he surmised that election officials:

Limited research thus far has shown that voters often miss finding errors that are introduced on the paper rolls. i.e. voters often don't bother to verify the paper rolls, and even when they do, voters miss seeing any errors on it.