Internet Voting: Proceed Cautiously

By Kim Alexander and David Jefferson

Published May 16, 2000
Copyright San Jose Mercury News.

Internet voting has recently captured the interest of election professionals and political activists as an alternative to voting in person at the polls or by mail with an absentee ballot. At first thought, electronic voting over the Internet seems a natural and even exciting prospect, one that promises to serve more voters more conveniently than ever before, including underserved groups such as students, working people, the military, travelers and others who cannot get to their home polls conveniently on election day.

Proponents argue that e-voting is a revolution that might kindle renewed interest in voting, especially among the young, and help raise the shamefully low voter turnout that plagues American elections.

We, too, were initially enthusiastic about the prospects for Internet voting. But after studying the issue for nearly a year as members of California Secretary of State Bill Jones' Task Force on Internet Voting, we have come to recommend great caution in any move toward Internet voting.

The more ambitious forms of e-voting have several potential dangers, including serious security, privacy, legal and policy problems that do not yet have satisfactory solutions.

The Task Force on Internet Voting was composed of election professionals, security experts, political scientists and representatives of non-partisan voter organizations to study the possibility of e-voting in California. Its report, which can be found online at, makes the fundamental point that not all Internet voting systems are alike.

There are two broad categories that have to be clearly distinguished. Type I systems are county-controlled, with the computers and software used for voting under the full control of election officials. In this category would be Internet voting machines at traditional poll sites on election day, and also kiosks, similar to ATM machines, located at convenient places around the county and open to voters for days or weeks prior to election day, allowing many more locations and much more time for voting than we now have.

Type II systems are those where election officials do not control the machine used for voting. Systems that permit voting from home, school or office computers, or from remote computers around the world, would fall into this category.

The two types of electronic voting systems differ greatly, in the following ways:

a) Security: Both kinds of systems are subject to certain generic kinds of cyber-attacks against their servers, such as the denial-of-service attacks that so easily brought down Yahoo, E*Trade, CNN and other huge sites recently. But Type II systems are also subject to Trojan horse attacks against the computers used for voting, which could allow not only spying on people's votes, but even changing them without detection. Similar attacks can be made through use of otherwise perfectly legitimate remote management software such as PC-Anywhere, LapLink, or Timbuktu.

These latter problems arise from the fundamental openness of the PC software environment; they simply cannot be circumvented by ordinary security techniques such as encryption or biometric voter authentication methods. And while there are other techniques to avoid these problems, they are not readily available today for use by voters, and all would entail substantial development costs or voter inconvenience.

b) Privacy, coercion, vote-selling: Type II systems would permit voting in uncontrolled institutional settings, especially the workplace. Historically there have been serious problems with vote selling and vote coercion in such situations. Also, there would be no way to guarantee the voter's privacy, either physical or electronic; in fact, current law, which allows employers to monitor employee use of company computers and networks, would actually undermine voter privacy.

c) Equal access: With Type I systems counties would be responsible for fielding voting machines in such a way as to serve all segments of the population equally, and they can be held accountable for it by the courts. Type II systems, in contrast, clearly give an advantage to prosperous people who have ready access to the Internet from home or office, possibly further exaggerating the voter turnout gap between rich and poor that already exists.

Besides these concerns, there are a number of other difficult and unresolved procedural issues. How does a California citizen exercise his or her right to observe the counting of votes when the ballots are electronic? What kind of audit trail or recount procedures make sense for electronic ballots? How will we educate legislators, election officials, and the public about encryption and other Internet security measures needed to implement e-voting?

In a democracy, the legitimacy of government depends on the security, openness and fairness of elections. The challenges of Internet voting may be solved in time, and in the meantime, pursuing Type I (county-controlled) systems makes sense on an experimental basis.

However, we should delay any implementation of any Type II systems until such time as the difficult problems inherent in them can be solved.

It is also important to keep Internet voting in perspective. Voting convenience, important as it is, would not be the most significant application of the Internet to the processes of democracy. Far more important is its ability to bring detailed, timely, relevant information to the electorate, such as government records, campaign contribution and spending data, and authoritative facts about the issues. Voter engagement and education are the real keys to a revitalized democratic process.

Kim Alexander is president of the California Voter Foundation, a non-profit group advancing new technologies to improve democracy at David Jefferson is a Voter Foundation board member and senior researcher at Compaq Systems Research Center, Palo Alto. He can be reached via e-mail at

Copyright 2000, San Jose Mercury News.

This article originally appeared in the May 16, 2000 edition of the San Jose Mercury News, Opinion Section, (

Main Page

What's New




Contact Us

Support CVF

This page was first published on May 25, 2000 | Last updated on May 25, 2000
copyright 1994 - 1999, California Voter Foundation. All rights reserved.