TOPICS > Science

Electronic Voting

May 5, 2004 at 12:00 AM EDT


SPENCER MICHELS: As the November presidential election approaches, groups of protesters have sprung up in California and across the country to rally against new electronic touch-screen voting machines. The new devices were designed to eliminate the hanging chads on punch cards and the confusion that disenfranchised Florida voters in the 2000 presidential election.

Last week in Sacramento , California Secretary of State Kevin Shelley, who had already banned the use of punch cards, banned 14,000 of the new machines in four counties. He also decertified all touch- screen systems used in 10 additional counties until steps are taken to upgrade their security for the November presidential election.

KEVIN SHELLEY: We are acting boldly and responsibly to improve and secure these systems in time for November.

SPENCER MICHELS: Shelley’s announcement means several California counties will have to scramble to get ready for the November election. And it adds to the nationwide political and scientific controversy over electronic voting. More than 50 million Americans had been expected to vote this year on Direct Recording Electronic voting machines, or DRE’s, as they’re called.

With this system, the voter touches a box on a computer screen next to the name of the candidate selected. The vote is recorded on the computer’s memory. After the Florida debacle, Congress authorized $3.9 billion for new electronic machines and ordered the states using old technology to upgrade. But once they were in service, 10 states, including California , reported problems.

In the Super Tuesday March primary, several California counties, including two of the state’s largest, Alameda and San Diego , experienced software failures, causing polling places to open late. And some voters were turned away when machines wouldn’t work.

Secretary of State Shelley appointed a voting systems panel to look into those problems. The panel delivered a scathing report and accused the state’s largest supplier of DREs, Diebold Election Systems, of jeopardizing the primary by using uncertified software and marketing a voting system before it was fully functional.

MARC CARREL: They’ve been doing a bait and switch on software that has resulted in the disenfranchisement of voters in various counties, and that has resulted in a reduction in the confidence not only of DREs, but in voting, in general.

SPENCER MICHELS: Diebold’s president, Robert Urosevich, defended his company.

ROBERT UROSEVICH: I want to be crystal clear that these allegations in this report about Diebold’s deceiving are not true and are factually not supported.

SPENCER MICHELS: After reading the panel’s report, Secretary Shelley asked the state’s attorney general to look into possible civil and criminal charges against Diebold.

KEVIN SHELLEY: We must send a clear and compelling message to the rest of the industry — Don’t try to pull a fast one on the voters of California because there will be consequences if you do.

SPENCER MICHELS: Mark Radke is Diebold’s marketing director.

MARK RADKE: This is a new technology, and any time you change technology, you’re going to have a number of people that don’t like the change.

SPENCER MICHELS: But technology advocate Cindy Cohn says new technology is not an excuse for equipment failure.

CINDY COHN: I really protest the idea that every technology rolls out has glitches all the time. That’s true for mass marketed consumer technology. When you and I started using ATM’s, they didn’t glitch – they didn’t fail. The problem is they’re trying to roll out prototype machines that haven’t been fully tested that have been slapped together largely because there is this flood of federal money.

SPENCER MICHELS: Yolo County Clerk Freddie Oakley was worried from the start about the new touch screen machines and the decision by some local officials to buy them so fast.

FREDDIE OAKLEY: They have made appalling decisions about how to spend vast quantities of public money. They listen to the salespeople. They buy the products. Now their backs are against the wall. They’re defending bad decisions.

SPENCER MICHELS: In Alameda County encoders that format the electronic ballot for each voter failed so even supporters of electronic voting, like elections official Elaine Ginnold admit to the rush to buy new machines after the 2000 election caused problems.

ELAINE GINNOLD: It was done much too quickly, much too quickly.

SPENCER MICHELS: The result of that.

ELAINE GINNOLD: The result of that is what we’re seeing today. We’re seeing machines being decertified, we’re seeing encoders that are rushed into production without fully being tested and fully certified. We’re seeing voter distrust.

SPENCER MICHELS: Despite all that, Ginnold says the electronic voting machines can be trusted to tabulate votes accurately.

ELAINE GINNOLD: For a month before the election, we’re putting votes on them, test votes and test modes, and we are printing out the reports from them and we see that they’re absolutely accurate. They will even tell us if we’ve made a mistake.

SPENCER MICHELS: The debate has gone beyond politicians and election officials and has extended in to the computer science community.

MAN: And so, I believe that there are sound engineering solutions to the very real problems there are with DRE machines.

SPENCER MICHELS: Computer security experts gathered in Berkeley to debate touch-screen voting. Stanford’s David Dill says computer programs, including voting machines, will always have bugs.

DAVID DILL: We’ve been trying to solve the problems of program bugs in computer science for 50 years. We haven’t succeeded. Any program of any size has bugs.

SPENCER MICHELS: Dill says he worries about the security of voting machines, which he says are vulnerable to both unintentional flaws and deliberate tampering. A rogue programmer, for example, could fix a machine to record votes for the wrong candidate.

DAVID DILL: A voter can go to one of these things, enter a vote for Candidate A. If a vote is stored in the electronic memory for Candidate B, there’s no way for the voter to know that. The situation that worries me the most, because it’s easy and because it’s undetectable, is if an insider, who has access to the software before it’s installed on the machines, could put something malicious in it. That could be done in ways where you can’t detect it by testing, but it’s triggered during a real election, and you can’t detect by inspecting the software.

SPENCER MICHELS: But Michael Shamos of Carnegie Mellon says that tampering is detectable.

MICHAEL SHAMOS: There are any number of checks to determine whether something has been altered by even by a single bit, a single one or zero has been changed in the machine, and we can tell. For years– in fact, about six years — I’ve had a challenge posted on the Internet that there’s a $10,000 prize for anyone who can tamper undetectably with a DRE machine. No one has taken me up on this and no one will take me up on this because it’s not possible to do it.

SPENCER MICHELS: Shamos, who is an elections consultant for two states, says people trust computers with their lives and money, and touch-screen voting machines can be trusted, too.

MICHAEL SHAMOS: If things don’t come out the way they are predicted, everyone smells a rat. And if you smell a rat, you can impound the machine, and you can inspect it until the cows come home. And if there is some rogue programming in there, you’re going to find it.

SPENCER MICHELS: Scientists like Dill say that perhaps the simplest and most essential solution to reassure voters is to keep a paper trail. At the computer conference in California , Dill arranged a mock election, where conferees recorded their votes on a machine which he had programmed to give a false result.

The point was to show that without a voter verified paper trail; even experts couldn’t detect vote tampering on touch-screen DREs. California ‘s secretary of state has ordered that, by 2006, counties must attach a printer to touch-screen machines. That will provide a paper printout that a voter can inspect to verify that the machine recorded his vote correctly. Besides reassuring the voter, says David Dill, the paper ballots will allow for an audit.

DAVID DILL: If the results of the audit show there’s a discrepancy between paper ballots and electronic records, the paper ballots have to win. And under those circumstances we can be pretty sure that votes we’re counting are the votes that the voters intended to cast.

MICHAEL SHAMOS: What I object to is the requirement that there be voter verifiable paper trails. These paper ballot trail states want to make that paper, which is the most easily manipulatable record that we have, they want to make that the supreme one. And I think it’s an error to do it.

SPENCER MICHELS: Despite Shamos’ reservations, some voting machine companies are in the process of getting printers certified, and legislation is pending in the U.S. Senate that would require all voting systems in the nation to produce a paper receipt.

Meanwhile, the banned California machines will be replaced by an optical scan system which uses a paper ballot that is counted by computer. Although blind people say they have trouble using optical scan machines, 55 million Americans will use them this year — slightly more than will vote on the controversial electronic touch-screens.