The ACLU "interim" statement on DREs

From: Barbara Simons <simons_at_acm_dot_org>
Date: Mon Aug 02 2004 - 10:55:07 CDT

Since some people asked me about the ACLU and Barry Steinhardt, below is the
ACLU "interim" statement on DREs. It had been posted on the League of Women
Voters national website for several months. Though Barry refused to say
that this was an official ACLU website, nor would he say it was a draft, he
refused to ask the LWV to remove the statement from their website, using the
lame excuse that he didn't want to stifle speech. We finally got the LWV to
remove the statement, but it took a lot of work.

Meanwhile, I don't know what the current ACLU view is of this statement. To
the best of my knowledge, it has not been withdrawn officially, but who
knows.

I hope that people on this list will contact the national ACLU office and
any members of the ACLU leadership who you know on this issue.

>> ACLU Statement of Principles on Touch Screen (DRE)Voting Systems
 

The integrity of the voting process is fundamental to the operation of our
democracy. A major component of a valid electoral process is voting
technology that honestly and accurately counts every ballot. Because voting
technologies have always been susceptible to error, bias, and corruption, we
must remain vigilant about new technologies and insist that they maximize
the likelihood of recording what each voter intends, regardless of the
voter’s race, economic status, or geographical location. To this end, we
must require that voting machines be accessible to all voters by reducing
barriers to participation erected by language, physical disability, or
complexity. Because democracy also requires that the public have confidence
in the results of elections, we must ensure that voting technologies may not
be rigged in a way that would thwart the true will of the electorate.

Though the now discredited punch-card voting systems failed all these tests,
there is much debate about what voting systems should take their place.
Touch screen voting systems offer tremendous potential advantages, including
ease of use, accessibility to persons with disabilities, ready accommodation
of the needs of language minorities, and the voter’s ability to review and
correct ballots. However, computer security experts have raised serious
concerns about whether those machines are open to undetectable error,
tampering, or outright fraud. Moreover, recent experience with the use of
touch screen voting machines in California and Florida illustrate that these
machines are vulnerable to more prosaic problems such as getting the
machines up and running, inadequate training of poll workers and others
responsible for overseeing the use of the machines, and instances in which
manufacturers have not lived up to their representations concerning the
machines, e.g. the support for multiple languages.

The ACLU strongly supports the recommendation of computer experts that
digital voting technologies be subjected to the most rigorous testing and
certification procedures. This should include rigorous and public testing of
the software used in these systems.

The Voter-Verified Paper Ballot

Some computer experts have recommended the inclusion of a contemporaneously
created, “voter-verified” paper ballot that would become the ballot of
record in the case of a disputed election. The ACLU has serious reservations
about the both the effectiveness and practicality of this proposal for the
following reasons:

>>> 1. Election officials would resort to a “verified paper trail” only in the
>>> case a recount or contest, which a hacker can prevent or deter. In most
>>> jurisdictions, recounts are triggered only when an election is close. Thus,
>>> anyone savvy enough to hack into a digital system and alter election results
>>> would simply select a margin of victory big enough to prevent a recount or
>>> discourage a contest. In these jurisdictions, a competent hacker could block
>>> the review of any paper ballots. Even in those few jurisdictions, like
>>> California, which automatically conduct a recount of a small percentage of
>>> the ballots, a sophisticated fraud could thwart detection by corrupting the
>>> code for the paper printout. (See paragraph 2 below)
>>>
>>> 2. The voter-verified paper trail could be used by a sophisticated fraud to
>>> give voters a false sense of security that their vote was correctly tallied.
>>> For example, if the computer code is genuinely vulnerable to attack, a
>>> competent hacker could not only compromise it to make the machine record a
>>> fraudulent vote, but could also compromise the code that runs the printer,
>>> causing it to display the voter’s intent while the machine records the
>>> fraud.
>>>
>>> 3. There is no reason to assume that paper recounts are more accurate than
>>> DRE machine tabulations. Paper is notoriously difficult to handle and easy
>>> to manipulate. Counting the paper ballots generated by DREs would be subject
>>> to all of the historical problems associated with paper ballots, including
>>> human error, fraud, and mishandling.
>>>
>>> 4. The reliability of printers has never been systematically tested in
>>> conditions similar to those that exist in polling places on Election Day.
>>>
>>>
Recommendations

The ACLU believes that the voter verified paper ballots should be not be
employed until there has been a rigorous test of their reliability and
practicality in circumstances comparable to their use on Election Day. This
review should include a consideration of the possibility for human error and
fraud in handling these ballots.

In the interim, if DRE’s are to be employed in the 2004 election cycle:
>>> 1. The computer source code for all security critical functions of the
>>> machines should be subjected to thorough independent review. “Open Source
>>> Code”, which can be freely tested, is the best solution to the problem of
>>> computer software integrity. At a minimum, the full the code should be
>>> subjected to a review by an independent body and only open source code
>>> should be used for tabulating the results.
>>>
>>> 2. Rigorous physical security measures need to be instituted to insure that
>>> the machines and any associated paper ballots are not compromised.
>>>
>>> 3. Election officials need to be thoroughly trained in their use and the
>>> physical infrastructure necessary to insure their use, e.g. sufficient
>>> electrical wiring, needs to be assured.
>>>
>>> 4. The jurisdiction should have a permanent broad-based security task force
>>> or oversight body, representing all interested segments of the community, to
>>> evaluate the potential for fraud or error in voting systems and to address
>>> the new security challenges that will inevitably arise in the future. That
>>> task force should have complete access to the DRE code and conduct its own
>>> independent testing.
>>>
>>> 5. Election officials should select technology that gives them maximum
>>> flexibility in taking advantage of emerging technological innovations,
>>> including the incorporation of printer that will provide a voter-verified
>>> paper ballot for use with touch screen systems, if such technological
>>> innovations are shown to be feasible and to enhance the integrity of the
>>> voting process.

If an election jurisdiction chooses to employ optical scan or a method other
than DREs for its general balloting, it should:
>>> 1. Be required to have a sufficient number of DREs available to accommodate
>>> the needs of persons with disabilities and,
>>>
>>> 2. Employ systems that can accommodate the needs of language minorities.
>>>
Finally, the ACLU believes that the debate over the voter verified paper
ballot has obscured other important issues that bear greater scrutiny. We
believe there needs to be:
>>> 1. An evaluation of the new generation of touch screen machines only now
>>> being developed and, in some cases, used in nations such as Brazil and
>>> Australia, and those being developed in the US by non-partisan and
>>> non-profit institutions. The review should especially focus on those
>>> systems, e.g., the Australian system, which are based on open source code.
>>>
>>> 2. Thorough and independent studies of the error rates including “lost
>>> ballots” that fail to record, as well as over and under votes, of the two
>>> newer voting technologies most likely to be employed in 2004 election--
>>> touch screen and optical scan ballots.
>>>
>>> 3. An analysis of existing testing and certification procedures for all
>>> digitally-based voting technologies, including both touch screen and optical
>>> scan voting technologies, to determine whether current procedures assure the
>>> integrity and security of all hardware, software, and any associated paper
>>> ballots.

==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Tue Aug 31 23:17:01 2004

This archive was generated by hypermail 2.1.8 : Tue Aug 31 2004 - 23:17:22 CDT