By 2028, Bradley Tusk wants every American to be able to vote on their phones.

It's a lofty goal, and one that most cybersecurity experts scoff at. But it's a quest that the venture capitalist and former political insider continues to chip away at.

His nonprofit, Tusk Philanthropies, announced a $10 million grant program Thursday to fund the development of a new internet-based voting system that he says will aim to win over security skeptics, who have long been wary of votes being cast via digital networks rather than through the paper ballots or ATM-type machines that most Americans currently use.

NPR is the first to report on the announcement.

"My goal is to make it possible for every single person in this country to vote in every single election on their phone," Tusk said in an interview with NPR.

An effort years in the making

Tusk was Uber's first political adviser, and he is also a former staffer for Sen. Chuck Schumer, D-N.Y., and former New York City Mayor Michael Bloomberg.

He has already bankrolled a number of small-scale mobile-phone voting pilot projects across the U.S. over the past few years, in which voters with disabilities and Americans living abroad from a select few districts have been able to return their ballots digitally.

However, the vendors that conducted those pilots have faced heavy scrutiny for security flaws in their systems as well as for a general lack of transparency around their software, as the source code for the underlying technology has remained private.

Those criticisms have spurred Tusk to fund the development of a new open-source option.

After a review process involving roughly 25 applications, his organization settled on Assembly Voting, an elections technology company based in Denmark, and the OSET Institute, a U.S. nonprofit dedicated to election technology and research.

OSET will design the public-facing ballot-marking application, and Assembly will design the technology that will actually transmit the electronic ballot from a person's phone or device to an election official.

That transmission process will be end-to-end verifiable, says Jacob Gyldenkærne, CEO of Assembly Voting, meaning a voter will have a way to make sure their vote was recorded and counted correctly and was not tampered with in transit.

End-to-end verifiability is considered a prerequisite by some cybersecurity experts for any internet voting system, though Gyldenkærne says even with such verifiability, he expects a flood of questions about the security of whatever system Assembly Voting comes up with.

"We are very open to academic researchers, ethical hackers and the security community," says Gyldenkærne. "It's a massive project, and it's important to say it's a development project. ... We do not have the holy grail."

Election technology provided by Assembly has never been used by a state or local government in the United States.

Tusk is optimistic skeptics can be won over

Assembly will face an uphill battle to win over a cybersecurity community that has been dubious, if not appalled, by the idea of such a mobile voting system.

"There is a firm consensus in the cybersecurity community that mobile voting on a smartphone is a really stupid idea," Duncan Buell, a computer science professor at the University of South Carolina who specializes in election technology, said in an interview with NPR last year.

Greg Miller, of the OSET Institute, says he has traditionally agreed with that line of thinking. But he has been heartened by what he sees as a seriousness from Tusk to engage with the security community, including by adhering to upcoming recommendations from a working group focused on digital voting at the University of California, Berkeley and by working with the cybersecurity company Synack to vet grant applications.

"This sounds like the right approach," says Miller. "It's not a race to commercialization. It's a paced progress towards a system that everyone can take a good clear look at."

Tusk says he knows security-minded experts may not give their blessing at first, but he's optimistic that once the open-source system is ready for piloting and examination, currently scheduled for mid-2022, at least some skeptics can be won over.

He also acknowledges that the conspiracy theories around election technology in the 2020 election may present an obstacle. Many election administrators may hesitate to experiment with new technology that could bring on even more claims of fraud and hacking.

But he says that such claims stem from how polarized politics are in the U.S. currently and that higher-turnout elections and especially primaries could offer a solution.

"The last few years in some ways have made [this effort] harder," Tusk says. "They've also made it more necessary."