E-Voting

Henry Norr 26.AUG.04
It is "so hard to understand how an Assembly committee last week killed a bill that would have required paper printouts for electronic voting machines," the Chronicle complained in an August 20 editorial.*

A review of the editorial board's own work would have provided part of the explanation: their failure to publicize and endorse the bill in question before the committee vote certainly made it easier for the legislators to vote it down. And a review of coverage of the e-voting issue throughout the paper would have suggested an even more important piece of the puzzle: The Chron has failed to give its readers and the state's leaders any real education about why the choice of voting equipment matters.

The sad reality is that the nation is rushing headlong to adopt voting technologies that are neither reliable nor secure. The result could well be election outcomes determined not by the will of the majority, but by computer malfunctions or, worse, by deliberate digital manipulation - high-tech fraud carried out by corrupt officials, by malicious hackers, or by employees of the private companies that make the voting hardware and software.

Printouts checked by the voter and deposited in a ballot box - known in election lingo as a "voter-verified paper trail" - would go along way toward solving this problem: They would give voters a way to ensure that their own choices are recorded correctly, and they would provide officials the means to verify electronically-generated totals and carry out recounts in close races and cases of equipment failure.

Around the Bay Area and across the country, a growing movement of experts and other concerned citizens has carried these arguments into the public arena. More and more elected officials are recognizing their merit; others are resisting. But the Chron, it seems, can barely be bothered to cover the controversy.

Quantity and quality

So far this year, through August 24, the paper has by my count run 65 articles about Google's initial public offering and 101 about the Laci Peterson trial.** In just 15 days this month it produced 20 stories about alleged funny-money contributions to California Secretary of State Kevin Shelley. But over nearly nine months it has managed only 19 - including last week's editorial - on the e-voting debate.

Are 19 stories enough to do this issue justice? Other Bay Area papers don't seem to think so. Over the same time period the much-maligned Oakland Tribune has delivered exactly three times as many articles on the subject - 57 - while the San Jose Mercury News has run 58.***

The numbers alone don't get to the heart of the matter. The real problem is that the Chron's coverage hasn't given readers who rely on it the information they need to understand the importance and the complexities of the debate. The paper certainly has done nothing to advance the story by uncovering information not reported elsewhere. And on its editorial pages it has done little to challenge public officials and mobilize public opinion around this issue.

The 19 e-voting pieces in the Chron this year have included one decent and prominently featured overview of the debate - a piece by John Wildermuth that ran on the front page back on April 26 http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2004/04/26/MNGQS6B1K81.DTL&type=printableand one good "Open Forum" column - by Bank of America computer programmer Daniel Kohanski on May 12 http://sfgate.com/cgi-bin/article.cgi?file=/chronicle/a/2004/05/12/EDGDU6J7EJ1.DTL&type=printable.

The paper also ran a couple of stories that detailed the problems that occurred during the March 2 primary in Alameda County, where a component in Diebold electronic voting systems failed in almost 20 percent of the polling places and an undetermined number of voters were effectively disenfranchised.

Most of the Chron's e-voting stories, however, have been ho-hum, after-the-fact reports on steps taken by government officials - Secretary of State Shelley, his office's Voting Systems and Procedures Panel, and registrars in individual counties. On the whole, the coverage has conveyed the impression that the problem with electronic voting is simply that certain systems from some vendors have been prone to "glitches." That implies an easy solution: just "weed out the bad and save the good," as an April 29 editorial put it.

The missing dimension

Now, the demonstrated unreliability of e-voting technology is certainly a serious issue - far more serious than the word "glitch" suggests. In addition to preventing voters from voting, as happened in several counties in March, malfunctions in electronic voting systems have in several recent, documented cases (and who knows how many that weren't caught) caused votes for one candidate to be credited to another or simply to disappear into cyberspace, never to be recovered and counted. As VerifiedVoting.org - a web site that tracks these issues closely - puts it, "electronic miscounting of votes is no longer a theory - it's a fact." (For details, check out the VerifiedVoting site; the database of "mess-ups" collected by another activist group, VotersUnite! ; or the other sources listed in the resource guide accompanying this article .)

Beyond unreliability, however, there's a more profound problem with paperless voting that the Chronicle has scarcely broached: insecurity. Over the last few years - even as California counties and other jurisdictions all over the country have been rushing headlong to buy touch-screen machines - an overwhelming majority of computer scientists and security experts has come to the conclusion that paperless voting machines are not safe from tampering.

They have made the case on two levels:

First, detailed technical analyses of the hardware and software used in paperless systems on the market today have turned up numerous security vulnerabilities. Avi Rubin, technical director of the Information Security Institute at Johns Hopkins University and leader of one such study last year, said "We found some stunning, stunning flaws." One of his co-authors, Adam Stubblefield, said that with $100 worth of computer equipment, "practically anyone in the country - from a teenager on up - could produce smart cards that could allow someone to vote as many times as they like." A "red team" assembled by a company called RABA Technologies at the behest of the state of Maryland to test the security of machines it was buying concluded that hackers could easily gain complete control of the machines and that "A voter can be deceived into thinking he is voting for one candidate when, in fact, the software is recording the vote for another candidate." And numerous other studies have reached similar conclusions.

Second, the computer scientists have argued that all-electronic voting is by definition insecure. As a petition signed by nearly 2,000 of them puts it, "computerized voting equipment is inherently subject to programming error, equipment malfunction, and malicious tampering." Bugs or, worse yet, rogue code planted by a hacker or -conceivably - by an employee of the voting-equipment company could easily cause the machines to appear to do one thing while actually doing another - say, confirming a vote for Kerry on-screen but actually adding it to Bush's total.

Sure, if it happened every time, it would very likely be caught in testing. But any programmer worth her salt could write code that, say, shifted two percent of the votes - and only on the first Tuesday in November in years divisible by four. As for the "event logs" and "redundant storage" and other security measures built into these systems, as long as they're created with software, they can be tricked with software.

That's why the computer scientists - in league with election experts such as Kim Alexander of the California Voter Foundation and a growing army of voting-equipment activists - are demanding that voting equipment produce a voter-verifiable audit trail - basically, paper ballots voters can check for accuracy, then deposit in ballot boxes, so they can be used if not for the initial vote count, then at least for spot checks on the machines' accuracy and for recounts in close or disputed races. Without such a record to fall back on, there's simply no way to verify that the results electronic systems return actually reflect the preferences of the voters.

The Chron, however, has never really presented these arguments and the empirical evidence behind them to its readers. Aside from Kohanski's op-ed and a few short paragraphs in Wildermuth's feature - which were immediately "balanced" by glib dismissals from flacks for the voting-equipment industry and undermined by an unsupported reference to "conspiracy theorists run[ning] wild" - the paper has essentially ignored the security issue.

To be sure, the Chron's editorial board has endorsed a paper-trail requirement, not once but three times: in last week's editorial, in the April 29 one, and in another that appeared last December. In each case, however, it has supported its arguments only with vague references to "unreliable technology" and the need to "instill voter confidence." And without a clear understanding of the very real risks to democracy associated with paperless machines, the case just isn't that compelling. (The Chron's April 29 editorial included a sentence as nonsensical as it was ungrammatical: "Since the infamous Florida vote of November 2000, money and technology is fast replacing punch cards.")

In contrast, the Oakland Tribune and the Mercury News - among other papers - have given extensive coverage to the security debate. In the case of the Trib, reporter Ian Hoffman has produced a steady stream of detailed news stories on the subject; one of them, focusing on a recent analysis of flaws in systems made by Diebold Elections Systems, began "Hacking a Diebold touchscreen voting machine is astonishingly easy with a few tricks and busy poll workers. All it takes is a $1 plastic card with a memory chip, like the ones voters use, and the encryption keys and passwords for Diebold's software."

Backing up Hoffman's intrepid reporting, the Trib has run four op-eds on the security question and the need for paper ballots, including strong ones by Paul Krugman, the popular New York Times columnist, and by Avi Rubin, professor of computer science and technical director of the Information Security Institute at Johns Hopkins University, who led an earlier study exposing the weaknesses of Diebold's software. And while the Trib's own editorials on the subject have been as fuzzy as the Chron's, there have been six of them just this year, all of them reflecting considerable skepticism about electronic voting.

The Merc has also given its readers a thorough education on paperless voting. Reporter Elise Ackerman, who held down the beat until she transferred to Washington, D.C., in June and began writing about Abu Ghraib, Guantanamo, and such, did an excellent job. On Feb. 1, for example, the Sunday Merc carried a pair of first-rate features, bluntly headlined "Electronic Voting's Hidden Perils" and "Rigged Election Called Possible - Researchers Find Vulnerabilities," under her byline. Through the spring she continued to follow the issue closely, and the paper supplemented her work with frequent AP reports on developments in the controversy outside of California, as well as commentaries by security expert Rubin and the Merc's own tech columnist, Dan Gillmor. And the paper's editorial board has chimed in with no fewer than 10 editorials on the subject - each of them clearer and more forceful than anything that's appeared on the editorial page of the Chron or the Trib.

Bureaucratic safeguards?

Unable to refute the technical arguments of the computer scientists and other critics of paperless voting, the technology's defenders - the voting-equipment manufacturers and, even more vociferously, many state and local election officials - have begun to concede that the machines might not be completely secure in theory. But in the real world, they've argued, governments set up a variety of safeguards that ensure that no one could take advantage of any hypothetical vulnerabilities without detection. These protections are said to include federal and state standards for hardware and software, testing and certification by independent labs to ensure compliance with those standards, and careful "logic and accuracy" testing at the local level before each election.

But as journalists - at some papers - and critics have begun to take a closer look at these supposed protections, it has become clear that they are largely sham. The most dramatic evidence to that effect surfaced last November, when it emerged that Diebold had installed software that had never been certified by any outside authority on Alameda County's voting machines before the elections last October and November. That discovery led to further investigations, which showed in short order that uncertified software was used not only in Alameda, but in all 17 California counties that rely on Diebold equipment, and that four counties were also using a hardware model that had never been qualified by federal authorities, despite repeated assurances from Diebold that it had.

Not surprisingly, these revelations were headline news in the Merc, the Trib, and just about every other paper in the state. But not a word about any of them appeared in the Chronicle. If you depended solely on the Chron for your news, you wouldn't have known anything about this scandal until late April, when the state's Voting Systems and Procedures (VSP) Panel began holding public hearings on whether to ban Diebold equipment from the state because of the company's willful misconduct.

At that point - five months after the story began to break in other papers - Greg Lucas, the Chron's Sacramento bureau chief, finally picked it up. (So much for the paper's in-print declaration that Lucas would no longer cover state government after his wife was appointed deputy chief of staff for Gov. Schwarzenegger….)

Lucas did report on the direct outcome of the Diebold certification scandal: orders by Shelley to ban the unqualified hardware, to impose a battery of new security procedures on counties using other touchscreen systems, to ban county purchases of new electronic voting machines unless they are equipped to produce voter-verifiable paper ballots, and to refer the Diebold case to the Attorney General for criminal as well as civil prosecution. A few weeks later Lucas followed up with a report that these orders had raised the hackles of some county officials and that four counties were taking Shelley to court to overturn them.

But in July, when a federal court judge responded to that suit with a ruling upholding Shelley's authority -an opinion greeted with a congratulatory editorial in the Merc and heralded by Hoffman in the Trib as a "landmark" because it was "the nation's first legal affirmation of a state's ability to require a so-called 'voter-verified paper trail' … for electronic voting". - the Chron didn't even mention it.

Meanwhile, Hoffman, digging deeper into the Diebold case, obtained and wrote about some revealing internal memos from the company's law firm. Among other things they demonstrated that the installation of uncertified software wasn't some inadvertent slip - even as Diebold was doing it, its own lawyers were warning that it was putting itself in legal jeopardy. (In response to that story, Diebold's law firm sued Hoffman personally as well as the Trib's parent company. To their credit the defendants promptly filed a motion accusing the law firm of a "strategic lawsuit against public participation," or "SLAPP suit," aimed at impeding free speech and free press rights, and the Diebold law firm then withdrew its suit.)

Hoffman also began looking beyond Diebold to other major equipment manufacturers. In one memorable investigative piece, he reported that California officials don't actually know which systems in use in the state - from all vendors, not just Diebold - are certified because, incredibly enough, the state has only fragmentary records of what was certified between 1999 and the spring of 2003, a period of rapid evolution for computerized voting systems in California.

At the Merc, Ackerman undertook a broader look at the whole certification system, and what she found should be as shocking as the technical flaws the computer scientists have uncovered. First of all, the federal standards under which voting equipment was certified until this year were drawn up in the late 1980s - aeons ago when you're talking about high-tech gear - and barely touched on the question of security. Moreover, equipment is tested for compliance with these standards not by government labs or even by truly independent organizations, but by a handful of for-profit companies overseen by a private elections group that's financed in part by the voting-equipment manufacturers - the very companies that make the systems being tested.

Another recent Ackerman piece highlighted another dynamic that raises serious questions about the effectiveness of the oversight system: the "revolving door" relationship between state and county election officials and the companies that make the voting-equipment they buy. For example, former Secretary of State Bill Jones (now the Republican challenger to Sen. Barbara Boxer) took a $10,000-a-month consulting gig with one leading manufacturer (Sequoia) after he left office, and at least three of his former staffers and two of his predecessor's staffers now work for the three largest voting-machine companies, according to Ackerman. The story wasn't new - the Los Angeles Times did a front-page piece on the same subject last November - but it was interesting and relevant to the current controversy. Yet nothing like it has appeared in the Chron.

League leaders

One final example: Until recently the leadership of the national League of Women Voters, the 84-year-old good-government group, actively supported touchscreen voting and opposed requiring paper ballots on the grounds that such a rule would slow down adoption of a technology that could benefit the disabled and minorities in particular. That line didn't sit well with many of the League's 130,000 members, and when the organization convened in June for a national convention, some of them decided to put forth a resolution that would change it.

Though the convention was in Washington, D.C., the story was a natural for Bay Area papers, because it had local interest as well as appealing characters and genuine suspense. One of the grassroots insurgency's main organizers was a 74-year-old Oakland activist, Genevieve Katz, who had collected more than 700 signatures in support of the resolution. And Barbara Simons of Palo Alto, a well-known computer scientist and former president of the prestigious Association of Computer Machinery, had decided to support the campaign by running for president of the League on a pro-paper-trail platform, in opposition to the incumbent, an outspoken advocate of paperless voting.

In the end, after intense debate, the convention voted to support "voting systems and procedures that are secure, accurate, recountable and accessible." Vague as the wording was, the decision was widely taken as a repudiation of the leadership's previous position and a victory for the grassroots rebels. Simons withdrew from the race for the group's presidency.

The Trib ran just one story on all this, but the Merc went to town with it, running wire reports before and after the convention, two staff-written stories focusing on Simons' role, and some follow-up discussion in Gillmor's column.

And the Chron? You guessed it: not a word about the conference, the grassroots challenge, or the outcome.

Civic responsibility

The point of all this is not that the Trib and the Merc have done anything extraordinary; in fact, there are important dimensions of the story they have yet to explore. But at least they've done what you'd expect a major metropolitan paper to do when a story that has such profound implications and generates such strong feelings emerges: They put capable reporters on the beat, stayed on top of major developments in the story, looked for local angles, and backed up their news coverage with forceful commentary and a consistent editorial voice.

The surprise is the Chronicle. Phil Bronstein likes to talk about making it a world-class paper, or at least he did when I worked there. But when it comes to the e-voting issue, the paper's performance has been strictly bush-league; in fact, its regional competitors have put it to shame.

The Chron does deserve kudos for one thing, though: In June the Insight section printed a piece by investigative reporter Greg Palast on the 1 million votes of African-Americans that were not counted in 2000 because they were declared "spoiled." I think paperless voting is a genuine danger to what's left of our democracy, but odds are it won't do as much damage as plain old low-tech racism already has. That's a story that, to my knowledge, none of the mainstream media in this country has really taken on. Palast's piece was at least a start.

* The Chronicle hasn't reported it yet, but the bill whose defeat so puzzled its editorial board has returned from the dead, now as AB 2454. For further information, see http://calvoter.org/news/cvfnews/cvfnews082004.html.

** All of these totals include staff-written and wire-service news articles, editorials, and columns, but exclude letters, corrections, and articles that make only passing reference to the topic. It's quite possible that I have missed a few items that should have been counted, but the big picture is beyond dispute.

*** It might be argued that these striking differences in coverage simply reflect the different election technologies in use, and therefore the likely interest levels of readers, in the papers' core territories: Alameda and Santa Clara counties now rely on touchscreen voting machines that generate no paper trail for recount and verification, while San Francisco uses optical-scan systems, which - though not above suspicion - at least involve paper ballots that can be rechecked.

In reality, though, those differences don't seem to explain the contrasting editorial choices, because only a handful of the Trib and Merc stories have focused on debates in their own home counties - both have covered the topic mainly as the state- and nationwide issue it is.

Besides, if the Chronicle is downplaying the issue because the city and county of San Francisco don't use paperless equipment, that's the height of parochialism: If - as mounting evidence suggests - the paperless systems 30 or 40 percent of voters in California and around the country will use this fall are untrustworthy, that's a problem that ought to concern everyone.

Resource Guide on Electronic Voting

If you really want to follow the e-voting controversy, no daily paper provides adequate coverage. Fortunately, the Internet makes it easy to fill in the gaps. - Henry Norr

Web sites

VerifiedVoting.org
The online arm of the grassroots campaign started by Stanford professor David Dill to require a voter-verified paper trail for all votes. Includes background papers, FAQs, and a vast library of clippings from the press.

VotersUnite!
VotersUnite!, another activist group, also provides valuable resources on the e-voting issue, including a clipping library and a database of documented "mess-ups" attributable to the hardware or software of the six largest voting-equipment manufacturers.

Kim Alexander's Weblog
Kim Alexander, president of the California Voter Foundation, is a hero of the struggle against paperless voting; more than any other single individual, she deserves the credit for pushing Secretary of State Kevin Shelley to mandate a voter-verified paper trail by 2006 in all California voting. Her blog includes excerpts from and links to major stories on the e-voting issue, with a special focus on California.

BlackBoxVoting.com and BlackBoxVoting.org
Two sites associated with Bev Harris, a Seattle-area publicist who "has made herself public enemy No. 1 for voting machine manufacturers," according to the New York Times, by relentlessly exposing their dirty linen - including some of their hopelessly insecure voting software. Her book, also called "Black Box Voting," can be downloaded for free.

Kim Zetter's "Machine Politics" coverage for Wired News
Reporter Kim Zetter consistently provides smart coverage of the e-voting issue for Wired News.

Electionline.org
A non-partisan, non-advocacy web site covering not only the e-voting controversy, but also other election issues.

Recent overviews

Ronnie Dugger, "How They Could Steal the Election This Time" in The Nation, Aug. 16, 2004
Dugger, founding editor of the Texas Observer, wrote a remarkably prescient article about the risks of electronic voting for the New Yorker way back in 1988. Now, in The Nation, he brings the story up to date, concluding that the rush to computerized voting has brought us to the brink of a "crisis for the integrity of democracy."

David Dill, "Summary of the Problem with Electronic Voting"
An introduction to the issue that boils it down to two short and clear pages, written by the Stanford professor of computer science who has become a key leader in the battle for voter-verified paper trails.

Kim Alexander, "E-Voting Reform: The Voting Rights Struggle of Our Time"
In this seven-page speech, delivered at the Commonwealth Club in San Francisco on July 7, 2004, Kim Alexander of the California Voter Foundation surveys the whole e-voting issue.

Ellen Theisen, VotersUnite!, "Myth Breakers for Election Officials"
This comprehensive, 53-page review of legal and technical issues with electronic voting is aimed at election officials, but it could also be useful for activists and others who want to explore the problem in depth.

E-mail clipping service

If you want to receive just about every article on e-voting from the mainstream press in your inbox, send a note to resist@best.com.

Video

"Invisible Ballots," a video featuring interviews with computer scientists and activists who have led the fight over e-voting, provides a very accessible introduction to the issue. It's available on DVD or VHS for $24.50 at http://www.invisibleballots.com. Both include a 90-minute version; the DVD also has a 50-minute version.