Sunday, December 16, 2007

Ohio study finds all e-voting systems vulnerable

Well here's a surprise: every electronic voting system tested in a recent study conducted at the request of Ohio Secretary of State Jennifer Brunner were found to have some kind of vulnerability, "which could impact the integrity of elections." From a press release on her website:
The Evaluation & Validation of Election-Related Equipment, Standards & Testing report, known as EVEREST, is a comprehensive review of voting systems revealing startling findings on voting machines and systems used in Ohio and throughout the country. The Ohio study tested the systems for:

- risks to vote security,
- system performance, including load capacity,
- configuration to currently certified systems specifications, and
- operations and internal controls that could mitigate risk.

The $1.9 million study, paid for using federal funds, was structured to allow two teams of scientists, corporate and academic, to conduct parallel assessment of the security of the state’s three voting systems - Election Systems & Software (ES&S), Hart Intercivic and Premier Election Solutions (formerly Diebold) - in both voting and board of elections environments. Separate research was conducted on each voting system’s performance, configuration and operations and internal controls management. A bipartisan team of 12 election board directors and deputy directors advised the study and evaluated all reports, participating with the secretary in making recommendations for change.

While some tests to compromise voting systems took higher levels of sophistication, fairly simple techniques were often successfully deployed.
The study resulted in a list of recommendations being published and sent to the state's governor. While if implemented they're likely to only affect Ohio's electronic voting laws, the same problems almost certainly exist in every other state that uses some kind of electronic voting terminal. Among the recommendations made were:
  • Eliminating points of entry creating unnecessary voting system risk by moving to Central Counting of Ballots
  • Eliminating Use of Direct Recording Electronic (DREs) and Precinct-based Optical Scan Voting Machines that tabulate votes at polling locations
  • Utilizing the AutoMark voting machine for voters with disabilities (This machine “reads” the bar code on a blank ballot and acts solely as a ballot marking device, allowing voters, especially those with disabilities, to mark ballots with little or no assistance, preserving the secrecy of their ballot selections.)
  • Requiring all ballots be Optical Scan Ballots for central tabulation and effective voter verification
  • Maintaining “no fault” absentee voting while establishing Early (15 days prior to the election) and Election Day Vote Centers (of the size of 5 to 10 precincts), eliminating voting at individual precincts or polling places of less than 5 precincts
  • Requiring all Special Elections (issues only) held in August 2008 to be voted by mail (no in-person voting, except at the board of elections, for issue-only elections held in August 2008)
While it seems like such a simple problem on paper, it's clear that even after four years of development electronic voting terminals are no where near ready for prime-time usage. In my neck of the woods (Broward County, FL), they're ditching the remarkably mediocre touchscreen kiosks that we used in the 2004 and 2006 elections and going back to virtually foolproof (unless your an idiot) optical scanning systems.

I have nothing against self-service where it improves things. But security, validation and reliability issues have plagued touchscreen voting machines from the get-go. Until somebody's really willing to sit down and address all of the concerns, and provide a completely open, transparent hardware and software stack that can be peer-reviewed and modified as necessary, I don't see how touchscreen voting machines could be seen to provide benefits that outweigh their risks and costs.

Tags: , ,

No comments: