In the fall of 2010, the District of Columbia was preparing to do something bold: allow overseas voters to cast their ballots online. A few weeks ahead of the November general election, it conducted a mock internet election and invited the public to try and hack the system. Within a few days, computer scientists at the University of Michigan had gained near complete control of the election server.
The team took control of webcams mounted inside the server room that housed the pilot, used login information to match specific ballots to specific voters, and changed not just votes that had been cast, but also ones that would be. “There is little hope for protecting future ballots from this level of compromise, since the code that processes the ballots is itself suspect,” the team wrote in a follow-up paper.
Afterward, D.C. officials confirmed that they had failed to see the attacks in their intrusion detection system logs, didn’t detect their presence in the network equipment, and only realized what had happened after seeing the group’s calling card: the University of Michigan fight song playing on the “Thank You” page that appeared after voting.
Technology has improved significantly since 2010, but internet voting presents a unique challenge. With paper, voters can verify that their ballot is correct before they mail it or insert it into a scanner. Once that ballot is tabulated, there’s no way to connect it back to the voter. It is irretrievable. When you cast a vote electronically, how do you ensure that the ballot the election office receives is the same ballot that you submitted — while also maintaining anonymity, producing an independent paper trail, allowing for some way to audit the results, providing publicly verifiable evidence if errors are detected, and ensuring that candidates can contest the results?
“No currently known technology—including blockchain—is close to enabling mobile or Internet voting systems to simultaneously achieve of all these requirements,” researchers from MIT wrote in a 2021 peer-reviewed paper titled “Going from bad to worse: from Internet voting to blockchain voting.”
The year of the D.C. meltdown, Brooke Pinto wasn’t yet living in the District. But after becoming the youngest-ever D.C. councilmember in 2020, she began working on one of the most expansive mobile voting bills in the country.
Allowing any eligible voter to cast a ballot using an internet-connected device, Pinto has claimed, would make voting more accessible, especially for voters of color, those with disabilities, and older adults. It could also raise the chronically low turnout rates in the District: 28 percent for the 2020 primary and around 67 percent for that year’s general, a little lower than the national average.
Pinto had no formal experience in voting or technology, but she said constituents had talked about it during her campaign. According to emails obtained through a public records request, in July 2021 her brother introduced her to Bradley Tusk, a venture capitalist who made millions from stock and options received from his work at Uber.
Tusk has poured money into a nonprofit and lobbying apparatus dedicated to pushing internet voting nationwide. His organization, Mobile Voting, has already conducted 21 internet voting pilots in seven states, and Tusk Philanthropies, which also works to combat food insecurity, has registered lobbyists in at least 12 states and Washington, D.C. Last year, Tusk Philanthropies announced a $10 million grant program to fund the development of a new internet-based voting system, according to NPR, in the hopes that every American will have the option to vote electronically by 2028.
His crusade could not come at a more perilous time for American democracy. Trust in the system is historically low, disinformation around voting technology is especially rampant, and many lawmakers are already suspicious of private companies providing any kind of assistance to election offices, which continue to be chronically underfunded.
After Pinto signaled her interest in pushing internet voting, Tusk’s team, including his organization’s D.C. lobbyist, Max Brown of Group 360, commissioned a public poll on the issue, rallied support from prominent civil rights groups like the NAACP, connected Pinto with experts, and helped draft the legislation and an op-ed in support of it.
Tusk argues that, because only extreme idealogues vote in their party primaries, politicians have no incentive to compromise, fueling hyperpartisanship and gridlock. But if we made voting significantly more convenient, he says, it would dramatically increase turnout and make policy more representative. “This is the only way to change the political inputs,” he wrote in a blog post earlier this year. “This is the only way to stop mass shootings, to come up with solutions to problems ranging from immigration to opioids, health care to education, climate change to infrastructure.”
However, assuming the technology is perfectly secure, there’s a problem. The foundation of Tusk’s argument — that internet voting will “exponentially” increase turnout — is contradicted by mountains of evidence, including research he’s funded. Even one of Tusk’s own grantees has spoken out against the bill.
None of this has dissuaded Tusk. “Short of committing a crime,” he’s said, there’s “nothing unethical about anything I could possibly do to try to make mobile voting happen.”
Mobile Voting declined to grant interviews for this story, and much of the information about its programs is available only through open records requests (or is not public at all). And Tusk’s philanthropic and venture capital arms did not respond to a request for comment.
Even when organizations providing funding around elections do disclose, trust is hard to come by. In the run-up to the 2020 presidential election, Facebook CEO Mark Zuckerberg and his wife, Dr. Priscilla Chan, donated $400 million to the Center for Tech and Civic Life. The money was meant to help election offices respond to the pandemic, every eligible jurisdiction that applied for a grant received one, and the company later published a copy of its tax form listing all of the funds it had distributed. Still, Republicans accused the tech billionaire of conspiring to bribe election officials and steal votes, with one Republican congresswoman decrying that the “liberal non-profit group flooded left-leaning counties … with opaque private funding to increase voter turnout.”
Described by Fast Company as “Silicon Valley’s political savior,” Tusk served as communications director for Sen. Chuck Schumer, D-N.Y., from 2000 to 2002; was deputy governor of Illinois from 2003 to 2006; and managed Michael Bloomberg’s 2009 campaign for a third term as mayor of New York City.
In 2011, he became Uber’s first political adviser and worked until 2015 to expand ride-hailing in major cities across the country. When he left the company, he owned what some reports estimated as $100 million in stock. Over the past 15 years, he’s personally donated at least $359,000 to Democratic candidates (and at least $25,000 to Republican ones), according to FollowTheMoney.org.
In 2015, he founded Tusk Ventures, a venture capital fund that primarily invests in startups in highly regulated industries, like gambling, cannabis, and insurance. A year later, he launched Tusk Philanthropies, which supports Mobile Voting and mimics the playbook that Tusk developed while at Uber.
When discussing a major election nonprofit that advocates for handmarked paper ballots, Tusk said on his podcast, “You know what paper ballots got us? Bush v. Gore. You know what that got us? The Iraq War. You know what that got us? Hundreds of thousands of people killed for no reason. My voting technology’s never killed a single person. Theirs has killed hundreds of thousands.”
Tusk also frames opposition as proof that he’s succeeding. “It probably goes without saying that those who like things the way they are—in other words, every current politician, interest group, union, major donors, and anyone else who has no interest in making it easier for people to challenge their power—will raise a host of objections to mobile voting,” Tusk wrote in his 2018 book, “The Fixer.” Regardless of what opponents say about security or practicality, he suggests, they’re really just afraid of the competition.
And it also means growing more quickly than even proponents would advise. Chris Walker is the clerk for Jackson County, Oregon, one of the places that has partnered with Mobile Voting to offer internet voting for military and overseas voters. Walker loves the system and wishes the state would expand its use to other groups, including voters with disabilities, victims of natural disasters, first responders, and out-of-state voters. But it’s imperative that the process move slowly, she said recently.
“Should it be mainstream right now? Absolutely not, I don’t think we’re ready for that.”
In 1999, at the height of the dot-com boom, California’s secretary of state convened a task force to study the possibility of voting online. “We went in assuming that the idea was, ‘Well, what’s the best way to bring internet building to the people of California?’” says David Jefferson, who was chair of the technical issues committee and has advised five successive Californian secretaries of state on technology-related issues. “I realized that there were profound security problems,” he says. “Not only that, but there were not going to be any solutions to them. … They were unsolvable.”
Still, Mobile Voting is trying to solve them by developing its own digital absentee voting system. On its website, Mobile Voting says that the system is based on the requirements identified by the U.S. Vote Foundation, and it links to a 2015 report to which Jefferson contributed. He says he’s unimpressed by what he’s seen from Mobile Voting so far. “There is nothing on that site but prose — no technical specs at all — not even a cartoon architectural diagram,” he wrote in an email.
Beyond the dearth of technical details, he also raised the issue of who would own the system; have legal liability for its failures; maintain, operate, and upgrade it; and — perhaps most critically if it were to be adopted nationwide — certify it.
“There is nothing on that site but prose — no technical specs at all.”
Jefferson has worked for years to stop internet voting legislation, including a bill in California that passed the Senate in May but was later pulled down by its author. Nationwide, at least 21 internet voting-related bills were introduced this year, according to Voting Rights Lab, which documents and analyzes voting and election laws. That includes the one in D.C.
In January 2022, Councilmember Brooke Pinto’s staff started editing the D.C. internet voting bill with Jocelyn Bucaro, a former elections official who now serves as the director of the Mobile Voting project, according to the emails obtained via the Freedom of Information Act. Over the next few weeks, the two parties exchanged at least three drafts of the legislation, according to emails obtained through FOIA. Meanwhile, Pinto also received guidance from Max Brown, the $17,000-average-a-month lobbyist working on behalf of Mobile Voting, according to lobbyist reports from D.C.’s Board of Ethics and Government Accountability.
To offer feedback on the bill’s security components, Brown emailed Pinto’s legislative director with the names of three people he described as “3rd party folks unaffiliated with us” and “arms length from our advocates. One was on Mobile Voting’s Circle of Advisors. (Brown says he was told they weren’t affiliated with Mobile Voting and that he never spoke to or met any of them).
Another was Andre McGregor, a former partner at ShiftState Security. In late 2018, ShiftState conducted a security assessment of Voatz, one of the internet voting vendors that Mobile Voting has hired to provide the technology for at least one of its pilots. McGregor told Slate that the company’s technology did “very well” but did not release the company’s underlying report.
Researchers at MIT later discovered that Voatz had “vulnerabilities that allow different kinds of adversaries to alter, stop, or expose a user’s vote,” and Oregon Sen. Ron Wyden asked McGregor to explain his positive review. “Several state officials have cited your company’s audit in response to my office’s inquiries about Voatz’s security or lack thereof,” wrote Wyden. “These officials believe, reasonably so, that Voatz’s technology passed a comprehensive audit.”
ShiftState did not release the underlying report. McGregor had been a part of Mobile Voting’s Circle of Advisors, but he says his involvement with the group ended in late 2021, before Brown sent that email to Pinto’s office. After being contacted by The Intercept, McGregor says he asked Mobile Voting to remove his name from its Circle of Advisors.
The third recommendation, Donald Kersey, is general counsel to another Circle of Advisors member, West Virginia Secretary of State Mac Warner. In his email to Pinto’s legislative director, Kersey criticized the bill for being “light on procedures,” like minimum security standards, and attached spreadsheets of technical materials. However, the bill requires that the Board of Elections be the one to develop the system’s security protocols — and within 180 days of the law taking effect.
“When people ask for best practices for voting online, it’s rather like asking for best practices for driving drunk.”
Ultimately, though, the bill’s lack of specificity may not matter, says Ron Rivest, a cryptographer at MIT who co-invented one of the most widely used algorithms to securely transmit data. “When people ask for best practices for voting online, it’s rather like asking for best practices for driving drunk,” he says.
Though Pinto has cited the example of online banking to suggest that online voting is also feasible, she acknowledged to The Intercept that “there are differences, of course, between the two.” She mentioned that banks have protocols for when an account is hacked and that the stakes are different when it comes to our democracy.
There are other challenges, including remotely verifying a voter’s identity and ensuring that a voter’s device is free from malware. But perhaps the biggest challenge that Pinto failed to mention is this: Unlike applying for a passport or submitting your taxes, casting a ballot must be done anonymously — and for that reason, it can never be recovered or modified. (Even in a blockchain-based electronic voting system, which allows anonymity and security, users must still use potentially vulnerable devices and network infrastructure, meaning such a system is still susceptible to serious failures.)
“Electronic ballot return faces significant security risks to the confidentiality, integrity, and availability of voted ballots. These risks can ultimately affect the tabulation and results, and can occur at scale,” says a 2020 report from the Cybersecurity and Infrastructure Security Agency, the National Institute of Standards and Technology, the Election Assistance Commission, and the FBI.
Even a grant recipient of Tusk Philanthropies doesn’t believe the practice is safe. “For the foreseeable future iVoting solutions introduce far more risk than benefit because there remain too many technical problems to verifiably solve,” said Gregory Miller, COO and co-founder of the Open Source Election Technology Institute, or OSET, which Miller said received a two-year, $1 million grant from Tusk Philanthropies.
On its website, Mobile Voting says that OSET was awarded funding to support the development of “an open-source, end-to-end verifiable mobile voting solution for digital absentee voting.” Miller says this is not a correct characterization.
“We do not consider ourselves as a ‘partner’ in their pursuit of mobile voting,” he wrote in an email. He added, “We are not actively researching or attempting to develop any software solutions to attempt to address some of the challenging problems to iVoting.”
He then linked to an open letter he says that OSET had signed, urging the D.C. Council not to pass Pinto’s bill.
On February 18, 2022, Pinto formally introduced the Mobile VOTE Act along with seven co-sponsors, a majority of the council. Its coalition of supporters would eventually include more than a dozen progressive organizations, faith groups, and racial justice nonprofits, including the D.C. chapter of the NAACP.
In April, the group demanded a public hearing for the bill, something that the chair of the committee in charge of the bill, Charles Allen, had refused to do, citing security concerns raised by experts. For Tusk, the lines were clear: Supporters of his bill, generally “people of color who really want to be able to increase access to voting,” he said on his podcast, and Allen, a “rich, white … out-of-district” member who represents “rich white people.”
Pinto, the bill’s champion, had lived in D.C. for eight years, did not vote in the District before 2020, and had only registered the year before. She was the only candidate in her ward to decline participating in the public financing program, personally contributed $45,000 to her campaign, and had the most money coming from out of state of all the candidates in her race.
During her campaign, she also boasted of being “the only candidate to have endorsements from a sitting U.S. senator and sitting U.S. congressmember,” both of which her family had reportedly donated to over the past decade. Critically, she was also endorsed by the Washington Post. “Unlike some candidates promising the sky under the banner of progressive justice,” the editorial board wrote, “she is steeped in reality and would hit the ground running with grit and smarts.”
In the 2020 primary, she won by only 379 votes — and with just 28 percent of the total. A little more than a quarter of voters turned out.
To put pressure on Allen, Pinto, Mobile Voting, and its allies framed the issue as a moral imperative. “If passed and signed into law, this expansion of voting options would represent a remarkable victory in the civil rights struggle of our day,” said Pinto and the Rev. H. Lionel Edmonds, senior pastor of the Mt. Lebanon Baptist Church, in an op-ed written in consultation with Tusk’s lobbyist, Max Brown, according to public records.
It appeared that the public was on their side. As at least one press clipping repeated, 70 percent of D.C. residents supported a mobile voting option. Mobile Voting often touts these kinds of customer satisfaction surveys, like the 100 percent of survey respondents in Denver who said that of all the methods of voting, they preferred voting on their smartphone.
“Unless you are seriously informed about the security issues, who would say no?” says David Jefferson. “It does sound like a good idea.”
Tellingly, another study found that West Virginia voters who were told that the system was secured by blockchain technology were less willing to use it again. “Maybe it tips them off to this idea that, ‘Oh right, there’s security risks here with mobile voting,” said the study’s author, Anthony Fowler, a professor in the Harris School of Public Policy at the University of Chicago.
To further vouch for the security of its pilots, Mobile Voting has also cited at least one “independent” audit conducted by the National Cybersecurity Center, a Colorado-based nonprofit established in 2016 “for cyber innovation and awareness.”
Mobile Voting has hired the NCC to manage and implement at least two of its pilot programs and paid them a fee to do so, according to memoranda of understanding obtained by open records requests. That includes a 2020 pilot in King County, Washington, which demonstrated that “with top-notch platform development, effective election official training and voter education, mobile voting can be accomplished securely,” said Forrest Senti, then director of business and government initiatives at NCC and now vice president of programs and operations — and another member of Mobile Voting’s Circle of Advisors. (The NCC did not grant an interview for this story and does not disclose its conflict of interest policy or financial statements).
In 2020, Tusk donated $40,000 to Shemia Fagan’s campaign to become Oregon’s secretary of state — the most he’s ever donated to a candidate and Fagan’s largest individual donor, according to FollowTheMoney.org. After Fagan won her race and became the state’s chief election official, a bill was introduced in the state’s legislature to create an internet voting system. The content of the legislation matched a document that Mobile Voting’s lobbyist had emailed to Pinto’s legislative director a few months earlier — and it would’ve put Fagan in charge of creating the rules around the system’s implementation.
Fagan did not support the bill, and the bill did not pass. Fagan told The Intercept her independent judgment was never compromised by Tusk’s donation and that she declined Mobile Voting’s invitation to join its Circle of Advisors, saying she wanted to focus instead on restoring trust in vote by mail. As for her position on internet voting, she says she did meet with representatives from Mobile Voting. “Ultimately, they could not refute the strongest concerns raised by the opponents,” she said.
In the end, Mobile Voting’s D.C. coalition wasn’t able to persuade Charles Allen, who never granted the bill a public hearing. But Tusk wasn’t daunted. “I haven’t started putting pressure on this guy,” he said in a podcast episode released in June. “There’s stuff that I’ve found out about him that hasn’t become public yet that’s gonna really fuck up his life.”
In a public survey Mobile Voting had commissioned about voting in the District, just 2 percent of respondents believed voting was difficult to begin with.
The District automatically registers voters, allows online and same-day registration, and could soon verify residents to vote without them having to register at all. For this year’s midterms, the Board of Elections has automatically mailed every actively registered voter an absentee ballot, which can be returned up to seven days after Election Day, one of the most permissive deadlines in the country. And the District offers eight days of early voting.
Nationally, D.C. is better than most, but it’s not as much of an outlier as you might expect. Forty-six states and D.C. offer early in-person voting, and 35 states and D.C. either conduct all-mail elections or offer no-excuse absentee voting.
“Convenience does help, but it doesn’t have a transformative effect” on turnout.
“Convenience does help, but it doesn’t have a transformative effect” on turnout, says Donald P. Green, a political scientist at Columbia University and the co-author of a book on turnout.
Paul Gronke, a political science professor at Reed College, agrees and points out that turnout is even more complicated for primaries, which Tusk considers so fundamental to his reform because of how heavily gerrymandered most districts are. “You don’t have party cues anymore, so you have to search for other cues,” Gronke says. That process takes substantially more effort and information.
Another underlying issue, says Fowler, the University of Chicago professor, is that most people don’t vote based on where their ballot matters most. If that were the case, turnout rates would be highest in local elections.
Tusk also argues that radical primary voters have hijacked our politics, which isn’t well supported by evidence, says Lee Drutman, a lecturer at the Johns Hopkins University Center for Advanced Governmental Studies. “Primary voters are more politically engaged and stronger partisans, but not significantly more ideologically extreme,” writes Drutman. Other researchers have found the same.
Nor is there compelling evidence that engaging more voters would significantly change our politics, according to a 2020 Knight Foundation survey of 12,000 chronic non-voters. “If they all voted in 2020,” the report said, “non-voters would add an almost equal share of votes to Democratic and Republican candidates.”
However, this group was less likely to actively seek out news, agree that votes are counted fully and accurately, or believe that politicians’ decisions have a strong impact on their lives — attitudes that internet voting is unlikely to change. Likewise, Tusk’s reform does little to register voters, combat gerrymandering, open up primaries, make elections more competitive, improve the quality of the candidates running, or diminish the power of dark money and political action committees.
As to the effect of internet voting on participation, would it increase turnout in the District — or anywhere else — by 10, 20, 30 percent?
“There’s zero evidence to support that kind of a claim,” says Fowler, who received a grant from Tusk Philanthropies for his research, which found between a 3 to 5 percentage point turnout bump in West Virginia’s internet voting pilot.
Internet voting’s potential to raise turnout is even weaker for the 15 percent of D.C. residents who don’t have smartphones. In “The Fixer,” though, Tusk makes a suggestion: “It’d be far cheaper to just buy them for everyone than to run a government this inefficient and this ineffective because no one bothers to vote.”
In theory, however, internet voting could substantially help other groups, something that Tusk uses to his advantage.
“We’ve either made it available to deployed military or people with disabilities,” Tusk said on a podcast in 2021. “We’ve found one group on the right that no one can object to, and one group on the left that no one can object to.”
Per a 2009 federal law, military and overseas voters have the option to receive a blank ballot electronically for federal elections, and 31 states also allow them to return that ballot electronically. Some states require that voters explicitly waive their right to a secret ballot.
In that light, the president of Democracy Live, one of Mobile Voting’s technology vendors, says he sees his company’s online portal not as a silver bullet, but rather as a better solution than the fax or email option commonly offered. “It’s not, ‘Let’s go do online voting for America,’” says Bryan Finney. Instead, it’s “Can we do that better? Can we do that more securely?”
Combined, the people casting ballots electronically represented less than .2 percent of all the votes cast in 2020. However, voters with disabilities are increasingly advocating to use those digital options for themselves.
These voters have always faced significant barriers. It can be difficult just getting to poll sites, which then present challenges of their own, and voting at home can be just as fraught, especially as states have limited the kind of assistance that voters can receive. Until recently in Indiana, voters could only be helped by a “traveling board” of elections officials who would come to their homes. Thanks to a recent court decision, they can now be assisted by almost anyone — but that’s not the victory it may seem, says Michelle Bishop, voter access and engagement manager at the National Disability Rights Network.
“We are basically telling them to ask someone else to mark a ballot they will never be able to verify and just trust that it was marked as intended,” she says.
For now, Pinto has backed off of internet voting. “Since introducing the bill, I have had many more conversations with residents and experts and my staff and I have read additional reporting on the issue,” she told The Intercept. “At this time, mobile voting is not ripe to move forward as additional security protections are likely needed to be considered.”
Still, others, if not Tusk’s, will inevitably pop up elsewhere, says MIT’s Ron Rivest. “It’s like Whac-A-Mole.”
And once a right is expanded to one group, another usually follows. In 2018, West Virginia officials said they had no plans to extend mobile voting beyond the relatively small overseas population. “Secretary [of State Mac] Warner has never and will never advocate that this is a solution for mainstream voting,” his deputy chief of staff told the Washington Post. In 2021, internet voting in the state was offered to voters with qualifying disabilities. This February, it was expanded again to certain first responders.
Or, as Tusk put it in 2019, “What we learned at Uber is once the genie is out of the bottle, it can’t be put back in.”
This reporting was made possible through grants from the Gumshoe Group and the Fund for Investigative Journalism.