Thursday, December 1, 2016

What 6 Top Election Experts Are Saying About the Next Big Step in the 2016 Recount Wisconsin judge likes hand counts, but won't order it under state law.

A half-dozen of the nation's top computer scientists who have spent years analyzing how electronic voting machines can malfunction or be hacked to alter vote counts have filed affidavits in the Wisconsin presidential recount in support of hand-counting the ballots.
The scientists have collectively spent decades advising top state election officials and the federal government’s election oversight panels on computer security and preventing inaccurate or tampered vote counts. Their affidavits support the idea that the Wisconsin Elections Commission should order the state’s counties to recount all paper ballots by hand, instead of leaving it up to counties to decide if they will do that or run them through high-speed scanners as on Election Day. They also make the case that hostile foreign governments like Russia are quite capable of hacking into state election systems, inserting malware to a range of components to alter the reported vote counts and then disappearing with no trace.

These assertions are unprecedented in a federal election recount and strongly bolster the Green Party’s argument that the states conducting a presidential recount should conduct the most comprehensive recount possible, starting with hand-counting paper-based elements in voting systems. In Wisconsin, that includes ballots and tapes from touchscreen machines.
A Wisconsin court began hearing testimony Tuesday afternoon on the Green Party's suit to force a manual recount. The court ruled against the Green Party, with Dane County circuit judge Valerie Bailey-Rihn saying the Greens fell short of the legal requirements to order a hand count, even though the judge personally thinks a hand count is a good idea.
"I follow the law. That's who I am despite my personal opinions," said Bailey-Rihn. "It's (the counties’) decision. It's their discretion. I may disagree with it… but I must follow the law."

What follows are excerpts from the court filings by the six computer scientists supporting a manual recount.

1. Harry Hursti. In 2005, Hursti, who is based in Finland, developed a series of tests showing that the vote count by Diebold voting machines could be altered. He co-authored a comprehensive study of Ohio’s voting machine vulnerabilities, participated in Princeton studies showing other widely used voting machines were similarly vulnerable to hackers, and studied these issues in Europe.

Wisconsin uses the same optical scanning systems that studies he conducted in other states have found are vulnerable, he said, saying the computer memory cards, software, tabulation systems and other elements can all be accessed. “Optical scan machines can be hacked in a manner that changes election results, and such an attack would likely go undetected during normal pre- and post-election testing,” he wrote. “If the scanners are hacked, using them as part of the recount process is likely to result in the same fraudulent election outcome. The only reliable way to detect attacks on the scanners is to recount the paper ballots by hand and compare the results to the electronic tallies.”

2. Douglas W. Jones. Jones, an associate professor of computer science at the University of Iowa, has been studying electronic voting systems since 1994. He has served on numerous federal panels and advised Congress and states about the technicalities of conducting reliable, verifiable and transparent elections.

His affidavit said optical scanning systems reading ballots routinely make errors. “No optical scan technology, including that used in Wisconsin, is capable of perfectly uniform and reliable scanning and electronic tabulation of voter marked ballots… The potential for different interpretations by genuinely impartial scanners is even greater when ballots are initially scanned one machine and recounted on another.” He went on to describe, “for various reasons,” why machines can’t correct these problems, but that a hand count of the ballots could.

3. Ronald L. Rivest. Rivest has been at the Massachusetts Institute of Technology since 1974 and teaches computer programming, algorithms, cryptography, theoretical computer science, network and computer security, and elections and voting technology. He invented many of the encryption keys and protocols that are now used in internet browsing and commerce.

“We have learned the hard way that almost any computer system can be broken into by a sufficiently determined, skillful and persistent adversary. There is nothing special about voting systems that magically provides protection against attack,” he said, saying that an attacker has many opportunities to place malware into the software codes that run these systems. “This malware may be set to be triggered when a particular event occurs—perhaps something based on the date, the jurisdiction, or the pattern of choices made in an early-cast ‘trigger vote.' The malware may be okay [as] dormant during so-called ‘logic and accuracy’ testing, only to be activated during the actual election.”

“While such an attack may naively seem unlikely or implausible, it is not the sort of attack that is beyond the resources of a powerful nation-state, and may be likely or plausible today depending on political circumstances,” he continued. “The ‘Stuxnet’ attack on Iranian nuclear facilities demonstrated that even converters that are not connected to the internet may be successfully attacked.”

4. Philip B. Stark. Stark is a UC Berkeley professor and dean and director of several mathematics and statistical computing programs and research institutes. He serves on the editorial board of the Journal of Election Technology and Systems and has consulted with numerous federal agencies and congressional committees. He was a member of a California commission in 2007 that ended up phasing out paperless voting systems in the state.

“As of this writing, the margin between the president-elect and the second-place candidate in Wisconsin is 22,525 votes in more than 2,939,490 ballots cast. Hence, errors in the interpretation or tabulation of less than 0.38 percent of the ballots could have caused a tie to appear to be a win,” he wrote. “To determine if the reported winner actually won requires verifying the results as accurately as possible, which in turn requires manually examining the underlying paper record—not merely rescanning and retabulating the ballots.”

“When the margins are small, as they are in the 2016 presidential election in many jurisdictions including Wisconsin, the amount of error required to alter the outcome can easily be less than the error that an optical scan system makes in inferring and tabulating voter intent from the ballots or other paper record,” Stark said, underscoring why a hand count is necessary.

5. Poori L. Vora. Vora is a George Washington University professor of computer science who has spent the past dozen years focusing on computer security, privacy and secure electronic voting machines.

“Software-based voting systems are very complex and may consist of hundreds of thousands of lines of code. It is not possible to find all bugs in voting system software; nor is it possible to completely characterize its behavior in all possible scenarios,” Vora said. “For the same reasons, it is not possible to determine with certainty the absence of malicious software hiding within what might appear to be many thousands of lines of legitimate software code. Additionally, it is not possible to confirm with certainty that the code running on the machines is the code that was examined.”

Vora said the cashier-like paper printouts on touchscreen voting machines that were printed after each voter cast their ballot needed to be examined and compared to the overall vote totals compiled elsewhere in the system. Rescanning the ballots would not necessarily produce a more accurate count, Vora said, but manually examining the paper record trail would.

6. Dan S. Wallach. Wallach is a computer science professor at Rice University who has studied electronic voting systems for more than a dozen years, since Congress passed the Help America Vote Act in 2002, which brought the latest generation of machines to all 50 states.

“My main message is that our election systems face credible cyber-threats generally, and in this election year those threats are magnified in light of the persuasive evidence of state-sponsored attacks against our elections,” he wrote. “Recounts and audits, particularly in tight races, are appropriate measures to take against these threats.”

“Foreign nation-state actors, likely Russia, broke into DNC computers and relapsed documents for expressly partisan purposes,” Wallach continued, adding that they were caught trying to rig the results of a national election in Ukraine in 2014. “We must ask ourselves the same sort of questions that arise in any security analysis. Does the adversary have the means, motive and opportunity to have their desired effect, and do we have the necessary defenses and/or contingency plans to mitigate these threats?”

Wallach said the best way to go forward was a manual recount of all the paper records.

“By conducting manual tallies, a recount will produce a tally that more accurately represents the intent of Wisconsin voters than an electronic tally,” he concluded. “A manual truly is particularly necessary here given the concerning evidence of Russia-sponsored hacking and the vulnerabilities of our election machinery. Luckily, Wisconsin is a state that has paper records of each vote, which can be used to verify elections. I believe the only appropriate recount in this circumstance is one that manually tallies those paper records."

No comments: