Rebecca T. Mercuri is not exactly a techno-phobe. She has a doctorate in computer science. She's president of a New Jersey software company. "I've got an iMac and three laptops, and that's just at home," she says.
But for two years she has been warning anyone who will listen that election officials rushing to spend big bucks on computerized voting machines are courting catastrophe.
"The electronic voting systems have been created to a very lax standard," she says. "There are all sorts of security loopholes and potential problems. And states are spending billions on these machines, which will have a very short shelf life."
Mercuri is in a growing chorus of skeptics who say paperless voting systems might not be the answer to the problem exposed by the punch-card disaster of the 2000 presidential election in Florida.
The critics are not Luddites clinging to their typewriters and struggling with their VCRs. In Maryland, which has just agreed to spend $55.6 million for 11,000 Diebold touch-screen voting machines, a group of Johns Hopkins University researchers said last week that the popular electronic system might have serious security flaws.
Across the country, top computer scientists raise the same red flags.
"The bottom line is that all of the existing electronic voting machines are fundamentally flawed," says Peter Neumann, principal scientist at SRI International's Computer Science Laboratory in California. "The real problem is that the federal election system standards stink. They allow totally unsecure voting systems to be certified."
A coalition led by Stanford University computer scientist David L. Dill has declared that computerized voting machines are "inherently subject to programming error, equipment malfunction and malicious tampering."
More than 900 computer professionals have signed the resolution, which has been posted on the Web at verifiedvoting.org, saying touch-screen machines should not be used unless they also print a paper record that can be checked in case of problems.
Locally, Tom Iler, director of Baltimore County's Office of Information Technology, says the state's decision to adopt touch-screen machines in 2001 came over the strong objections of an advisory panel he served on.
"What if the original programmer put in an algorithm that changed every 30th Republican vote to Democratic?" Iler asks. "How would anybody know?"
Chads and butterflies
The drive to computerize voting began in the 1980s but got a huge push from Florida's hotly disputed presidential vote. After the embarrassing spectacle of partisan battles over "hanging chads" and "butterfly ballots," election officials and politicians were eager for change.
Naturally, they turned to the computer. And manufacturers of computerized voting equipment were happy to oblige.
"It's the lure of precision," says Clifford Stoll, an astronomer, cyberculture critic and author of High Tech Heretic: Reflections of a Computer Contrarian. "It's the lure of saying, 'We're high-tech. We don't want to be stodgy.'"
But as Stoll has documented with computers in schools, the assumption that computers were better than older alternatives was not closely examined. It should have been, many computer experts say.
Stephen Ansolabehere, a Massachusetts Institute of Technology political scientist and director of the CalTech/MIT Voting Technology Project, says there are good reasons to consider newer voting technologies. "Election administrators don't really like paper - a whole lot of ballots to print and a huge production to get ballots out to the polls and back in at the end of the day," he says.
But not all paper systems are as bad as Florida's punch cards. The Voting Technology Project compared the reliability of voting systems used nationwide from 1988 to 2000 and came to a remarkable conclusion.
"The most stunning thing in our work was that hand-counted paper ballots were better than anything else," Ansolabehere says.
Close behind were optical scanning machines that read ink marked on a paper ballot. Next came old-fashioned lever machines - the study found one in Tennessee installed in 1923 and still working fine. Lever machines were accurate at the top of the ballot but produced mistakes at the bottom.
Fourth in reliability came computerized voting machines, which edged out only the punch-card systems that proved so troublesome in Florida, Ansolabehere says.
While that study measured reliability, many experts are even more concerned about computerized systems' security.
"The ways that computer voting systems are manipulated are very hard to detect," Stoll says. "If a ballot box is lost, it's obvious. If a passel of electronic votes is lost, it's not obvious."
Report from Hopkins
That was the point of the report Thursday from Hopkins' Information Security Institute. It described what scientists said were security weaknesses in the software behind Diebold Election System's touch-screen voting machines, 50,000 of which have been installed nationwide.
The report, written by Aviel Rubin and colleagues at Hopkins and Rice University, concluded that hackers might be able to alter the software to allow multiple votes or vote-switching.
Diebold issued a statement Friday disputing the report but offering to "work directly with Johns Hopkins, its research team and other objective electronic voting experts to continue to ensure the integrity of the voting process."
Even as warnings about computerized voting have accumulated, federal and state officials have hurried to install the new equipment.
One strong selling point: The new machines often have an audio capability to help blind voters cast their ballots, and they can display multilingual text for immigrants.
Under the Help America Vote Act, Congress offered states $3.8 billion to revamp their election procedures. But a presidential commission to oversee the reforms, which was supposed to be in place in February, hasn't been appointed. Neither has a 14-member technical committee that's supposed to develop standards for the new machines, Mercuri says.
In Maryland, too, critics believe that the changes have been rushed. A governor's committee decided in February 2001 that a single voting system should be used statewide - a change most experts favor. But the group also dictated that the new system should use a computerized touch screen.
To decide which touch-screen model was best, the State Board of Elections convened a five-member panel to review different vendors' machines. But as they saw how untested the systems were, members became convinced that the technology was not ready for deployment, according to Baltimore County's Iler, who served on the panel.
On Oct. 24, 2001, the four outside members of the panel - the fifth was a state employee - wrote the elections board declining to endorse any of the touch-screen systems. But their report was ignored, Iler says.
In response to the Hopkins report, Linda H. Lamone, the state election administrator, said yesterday that Maryland's experience in the 2002 election gave her "absolute confidence" in the Diebold touch-screen system, already deployed in several counties.
She said the machines not only met state and federal standards but "passed the one certification process that matters most - an election."
The skeptics reply that while problems with computerized voting systems might not show up in every election, they could be devastating when they do appear.
"There hasn't been a big scandal, so I suspect the machines will continue to be deployed," says Edward Tenner, an historian of technology and research associate at the National Museum of American History. "The big question is whether in some future election, there will be something cataclysmic," such as an election outcome changed by computer glitches or deliberate hacking.
"It's a national security issue," Tenner says. "It could create real political instability."