Electronic pollbook security raises concerns going into 2024

“We have a trust issue in elections. The more we can say there are standards that equipment must be tested to, the better,” said Larry Norden, an election security expert with the Brennan Center for Justice. “It’s like a seal of approval that really doesn’t exist right now.”

Poll workers use electronic pollbooks to check in voters. They typically are a tablet or laptop computer that accesses an electronic list of registered voters with names, addresses and precinct information, with some doing so through an internet connection.

Testing standards and a certification program for voting machines have been in effect for years, a process overseen by the U.S. Election Assistance Commission. While compliance is voluntary, most states will use at least some aspect of the federal process to ensure their voting and ballot-counting machines are secure and functioning properly.

But there is a much wider system of technology that supports U.S. elections beyond the devices used to scan and tally votes — from electronic pollbooks to voter registration databases and systems used to report unofficial election results to the public. Their use has been expanding rapidly in recent years.

Nearly one-third of all voting jurisdictions in the U.S. used electronic pollbooks in 2020, compared with about 18% four years earlier, according to data collected by Election Assistance Commission.

The systems come with unique security challenges.

In 2016, Russian hackers scanned state voter registration systems looking for vulnerabilities and even accessed the voter registration database in Illinois, although an investigation later determined no voter data was manipulated. In 2020, Iranian hackers obtained confidential voter data and used it to send misleading emails to voters, seeking to spread misinformation and influence the election.

Experts say the systems could be prime targets again for those seeking to disrupt the voting process and sow chaos around U.S. elections. Gaining access to a voter registration database, for example, could allow someone to delete voters from the rolls. When people show up to vote, they are told they are not on the list.

Although those voters would be allowed to cast a provisional ballot that eventually could count, widespread problems with the voter registration database would trigger questions about a process that already has suffered a loss in public confidence following a sustained campaign by Trump and his allies to discredit the results of the 2020 presidential election. There is no evidence of widespread fraud or manipulation of voting equipment in 2020, backed up by exhaustive reviews in states lost by Trump.

In Detroit last November, a few polling locations had a brief issue checking in voters related to a data error that was quickly identified and resolved. Trump seized on the early reports, calling the situation in Detroit “REALLY BAD” in a social media post and urging people to “Protest, Protest, Protest!”

Unlike voting machines that are not directly connected to the internet, many electronic pollbook systems are connected by design. Some are quite sophisticated.

In counties that have put in place a vote center model, where registered voters can cast a ballot at any polling place, electronic pollbooks must be able to communicate with each other and with a central system. That’s to ensure voters are not able to cast ballots at multiple locations or vote in-person after returning a mail ballot.

While that can present significant security challenges, scrutiny for the pollbook systems is not as consistent as with voting machines.

The lack of national standards has left state and local election officials on their own. For the 2020 election, 15 states, including Arizona, Florida and Nevada, did not require any type of electronic pollbook testing or certification, according to federal data.

States and even some counties are often testing their pollbook systems in isolation and results are not routinely shared — an information gap that could be addressed with a national testing program.

“Having that type of knowledge allows them to put compensating controls into place, but they are doing it on an individual basis — state by state, county by county,” said Ryan Macias, an election and security expert who advises federal, state and local officials.

Aware of the risks, many election officials require back-up measures, such as paper copies of voter lists at polling locations. Election officials and experts note that one advantage of national testing standards for voting machines is the ability to assure voters that they have been properly scrutinized.

Two efforts are underway that seek to address the lack of uniform testing standards for electronic pollbooks. The Election Assistance Commission partnered with the nonprofit Center for Internet Security to test pollbooks and other nonvoting machine technology. But the federal agency began working on its pilot testing program in late 2021, about the same time the center announced results of the first phase of its own project.

It’s not clear why the two groups went their separate ways and what will happen next. A spokesman for the center, Jay Billington, said the group is “close to concluding the pilot” and expects to provide an update soon.

Thomas Hicks, chair of the commission, said the agency is making progress on its own pilot program, but that it was unlikely testing standards could be in place before the 2024 election.

“But this is why we move forward,” he said. “In 2026, there will be another federal election, and in 2028 another.”

Hicks said he welcomed the work done by the center and thought having more than one testing program could allow states to pick the best option for them.

Experts said having national testing standards would go a long way to reducing costs of the systems and lessen the burden on state and local election officials to navigate security on their own. Companies that make the equipment have expressed support for the effort.

During a November 2021 panel hosted by the commission to discuss its pilot project, representatives from testing laboratories said they had evaluated 76 different pollbooks by about a dozen manufacturers over the past three years. Agency officials noted the stakes were high.

“Real or perceived attacks on our voting systems can threaten voter confidence,” one commissioner, Don Palmer, said during the panel. “So that’s one reason why we think as much testing as possible is a good thing.”