US elections are systematically rigged through hacked or pre-programmed voting machines that secretly flip votes to change the outcome
Where the evidence lands: ContradictedThat US elections are systematically stolen through electronic voting machines that are hacked, connected to the internet, or pre-loaded with software designed to flip votes from one candidate to another or to fabricate votes, so that the announced results do not reflect how people actually voted, and that officials and voting-machine companies conceal this.
Believed by: A recurring claim raised by various people across the political spectrum and across multiple election cycles; after 2020 it was most visible on the American political right, though distrust of opaque voting technology has appeared on the left in earlier years as well
The full story
Two different claims, kept apart
This is one of those subjects where almost all the confusion comes from folding two separate claims into one. Keeping them apart is the whole task.
The first claim is documented and true: electronic voting machines are computers, computers can have security flaws, and independent researchers have shown, under controlled conditions, that specific machines can be tampered with. That work is legitimate. It is precisely why the United States leaned so hard, over the past two decades, into voter-verified paper ballots and post-election audits. Trust in an opaque system has to be earned, and honest scrutiny is how it gets earned.
The second claim is the one this file rates, and it is much larger: that some real US election was actually decided by secret machine manipulation, that votes were quietly flipped or invented to change who won, and that the truth is being hidden. That is the claim that surged after 2020, recurred in 2024, and is already circulating in anticipatory form ahead of the 2026 midterms. On the evidence, it is debunked.
One more thing at the outset, because it matters legally and factually: nothing here accuses any voting-machine company or any person of rigging an election. The finding is the reverse. When the claims were tested, against paper, against audits, and in court, the rigging was not there.
The strongest honest version
The suspicion deserves its best form, because parts of it are built on solid ground.
Election security is a real field, and the flaws are real. At public events like the DEF CON “Voting Village,” and in peer-reviewed work, researchers have taken apart specific machines and found genuine weaknesses. In 2022, CISA itself issued an advisory about vulnerabilities in one Dominion product line. None of that is invented. Anyone who says voting machines are simply unhackable is overstating the case, and the people who study these systems are the first to say so.
Opacity is a fair thing to distrust. A voter marks a choice and a computer records it somewhere the voter cannot see. Asking citizens to take that on faith, in the highest-stakes act of a democracy, sits uneasily with anyone of a skeptical temperament. This is not a partisan point; concern about paperless machines was voiced on the left in the 2000s and on the right in the 2020s.
Paper trails and audits exist because the machines cannot simply be trusted on their own word. That is not a concession to conspiracy theory. It is the reason the conspiracy theory can be answered.
So the honest core of the concern is this: opaque systems should not be trusted blindly, real vulnerabilities have been demonstrated, and a free society is right to demand that every count be checkable. Every word of that is defensible. The question is what it actually shows.
Where the outcome-altering claim breaks down
The demand that elections be checkable is not the conspiracy theory. The conspiracy theory is the specific answer, that machines secretly flipped the vote, supplied as though the demand had already been satisfied in its favor. It has not.
The decisive move is to notice the gap between a machine can be tampered with in a lab and this election was tampered with in reality. A demonstration with unlimited physical access, unlimited time, and no observers tells you about a possibility. An election runs under chain-of-custody rules, bipartisan observation, logic-and-accuracy testing before voting, and, crucially, a paper record the voter can verify. The lab result is a reason to keep those safeguards, not evidence that they failed.
And when the safeguards were exercised, the manipulation did not appear. This is the part that is hard for the theory to survive: the paper matched the machines. Where results were audited or recounted by hand against the ballots voters themselves marked, the totals lined up. A scheme that silently flipped votes inside the machines would be caught the moment humans counted the paper. Again and again, the hand counts confirmed the machine counts.
The supporting pillars fall the same way. Vulnerability is not exploitation: CISA published the flaws it found and stated it had no evidence they were ever used in an election. A viral glitch is not a scheme: the “caught on video” incidents traced to calibration issues or voter mistouches that officials corrected on the spot. Distrust is not fraud: that many people came to believe the claim, after it was broadcast to them, measures persuasion, not ballots.
What audits and courts actually found
Two independent kinds of check bear on this, and both point the same direction: the physical audits and the courtroom record.
On the audits, the clearest example is Georgia in 2020. Faced with a close presidential race, the state conducted a full statewide hand tally of roughly five million paper ballots as a risk-limiting audit. The hand count confirmed the machine-counted winner, with discrepancies well within normal human-counting error. Risk-limiting audits and recounts in other states matched their machine tallies too. Around 98 percent of votes cast in 2024 had a paper record behind them, and officials sample those records after the fact for exactly this purpose. The mechanism that would expose machine flipping was run, and it exposed nothing.
On the courts, defamation law provided an unusually direct test, because such cases turn on whether damaging statements were true. A Delaware judge ruled that the disputed on-air claims about Dominion were false, and Fox News settled Dominion's suit for 787.5 million dollarsin 2023, the largest known media defamation settlement in US history. Newsmax acknowledged the falsity of certain Smartmatic claims and settled; Smartmatic's case against Fox has continued. The through-line is that the specific, named machine-rigging accusations were found to have been made without evidence.
When the paper was counted, it matched. When the claims reached a courtroom, they were found false. Two different systems, checked two different ways, returned the same answer.
It bears repeating that a defamation finding is a ruling that the accusations were false, not a finding that any company committed fraud. The record clears the machines; it does not convict them.
Why the claim keeps coming back
Machine-fraud claims are not a single event but a recurring pattern, appearing in 2020, again in 2024, and pre-emptively before 2026. The durability says something about how the belief works.
It feeds on a real seed. Because authentic vulnerabilities exist and get reported, the theory always has a true headline to point to; it simply stretches a lab possibility over a real result. A conspiracy theory attached to something genuine is far stickier than one invented from nothing.
It answers an emotional need. A surprising or narrow loss is easier to bear if a hidden machine did it than if one's fellow citizens simply chose otherwise. That pull is not confined to any one party; the shape of “the count was rigged” has been reached for across the spectrum when results disappointed.
It is built to resist disproof. In the usual telling, an audit that confirms the result is folded into the plot, and any glitch is promoted to proof. When no possible evidence is allowed to count against a claim, the claim stops being an empirical question and becomes an identity, which is why refutation so rarely moves it.
And it is amplified for reach. In the hours after a contested vote, before any audit can finish, the boldest assertion wins the most attention, and prominent voices have repeated machine-rigging claims to very large audiences. By the time the paper is counted, the story is already everywhere.
Where the evidence lands
Hold the two claims apart to the end. Scrutinizing voting technology is legitimate and worth doing: the machines are opaque, real vulnerabilities have been found, and every result should be checkable against paper the voter verified. That is not the conspiracy theory; that is the reason the conspiracy theory can be answered.
The rated claim is the larger one, that machines secretly flipped or fabricated votes and changed who won a US election, and the truth is concealed. Against that claim the hand counts matched the machines, the federal and bipartisan findings reported no manipulation, the disclosed vulnerabilities were never shown to be exploited, and the courts found the specific named accusations false. On that claim the verdict is Debunked.
None of this accuses anyone. No company has been shown to have rigged a vote, and the honest security researchers who find flaws are doing valuable work, not exposing a theft. The line to hold is the one this case is really about: between legitimate election-security scrutiny, which strengthens democracy, and the false story that machines quietly decided elections, which corrodes it. Keep demanding that the count be checkable. When it is checked, believe what the paper says.
What's still unexplained
- Electronic voting machines, like all computers, will always have some vulnerabilities, and the security research that finds them is legitimate and ongoing. The open work is hardening equipment, expanding routine audits, and keeping robust paper trails, not because any election was flipped, but because public trust in an opaque system has to be continually earned.
- Not every US jurisdiction audits to the same standard, and a minority of places have historically used equipment with weaker paper records. Closing those gaps so that every result can be checked against voter-verified paper is a real and unfinished policy question, distinct from the false claim that manipulation has occurred.
- How officials and platforms should respond to anticipatory fraud claims, the versions that circulate before any votes are counted, is an unresolved problem of the information environment, and it will recur around the 2026 midterms regardless of what the machines actually do.
- This file rates the outcome-altering claim against the evidence available through mid-2026. Election security is a live field; genuine future findings should be judged on their merits, which is a different posture from assuming the theft in advance.
Point by point
The claim: Security researchers have hacked voting machines, which proves elections are being stolen by machine.
What the record shows: The first half is true and the conclusion does not follow. Independent researchers have demonstrated real vulnerabilities in specific machines, but almost always with direct, prolonged physical access under laboratory conditions that do not resemble a monitored polling place. A demonstration that a device can be tampered with by someone who is handed it in a lab is a reason to keep paper backups and run audits, which is exactly what US jurisdictions do. It is not evidence that any real election was actually altered, and no such real-world exploitation has been found.
The claim: Machines secretly flipped votes in a real election and changed who won.
What the record shows: This is the central claim, and it fails the most direct available test. Where results were checked against the paper the voters themselves marked, the paper matched the machines. Georgia hand-counted about five million presidential ballots in 2020 and confirmed the machine outcome; risk-limiting audits and recounts elsewhere did the same. If machines had silently flipped votes, a hand count of the paper would expose it. Repeatedly, it did not.
The claim: Federal authorities and election officials are covering up machine manipulation.
What the record shows: The public record runs the other way. CISA and the Election Infrastructure councils, alongside bipartisan state and local officials, stated there was no evidence any voting system changed votes or was compromised, and CISA even published the specific vulnerabilities it did find while noting they had never been exploited in an election. Agencies disclosing flaws and inviting audits is the opposite of concealment. No investigation has produced evidence of outcome-altering manipulation to conceal.
The claim: The named voting-machine companies rigged the vote, as widely broadcast in 2020.
What the record shows: The specific, named accusations were tested in court and did not hold. A judge found the disputed on-air claims about Dominion to be false, and Fox News settled Dominion's defamation case for 787.5 million dollars; Newsmax acknowledged falsity in and settled Smartmatic-related claims, and further litigation continues. Defamation cases turn precisely on whether damaging statements were true, and these established that the rigging accusations were made without evidence. That is a legal record that the claims were false, not proof that any company did anything wrong; no company has been shown to have rigged an election.
The claim: So many people distrust the machines that something must be wrong with them.
What the record shows: Widespread distrust is real, but it measures confidence, not fraud. Much of the distrust was manufactured by the very claims at issue here, amplified in the days after an election before any audit could be completed. Polls of belief are not audits of ballots. When the actual paper is counted and the actual code and procedures are examined, the manipulation the distrust assumes does not appear. A belief becoming popular is not the same as its becoming true.
The claim: Vote-flipping was caught on video at the polls, so it is happening.
What the record shows: The viral incidents that reach this description have generally been traced by election officials to ordinary causes: a voter mistouching a screen, a calibration issue on a single machine that is then corrected, or a misread of a printed ballot summary. In documented cases such as isolated 2024 reports, officials let the voter verify and cast the ballot they intended. A handful of correctable, individually resolved glitches, spread out of context, is not evidence of a system-wide scheme to change an outcome.
Timeline
- 2002-2004After the disputed 2000 election, the Help America Vote Act pushes states toward electronic voting. Early academic studies and some activists, at that stage often on the left, raise concerns about paperless touchscreen machines and the difficulty of auditing them, seeding a durable distrust of opaque voting technology.
- 2006-2016Security researchers repeatedly demonstrate, in controlled settings, that specific voting machines can be tampered with if an attacker has direct physical access. These findings drive a real, bipartisan policy shift toward voter-verified paper ballots and post-election audits. The lab results are genuine; they are about what is possible under contrived conditions, not evidence that any election was altered.
- 2020-11Following the US presidential election, a wave of claims asserts that voting machines, naming companies such as Dominion and Smartmatic, switched or deleted votes to change the outcome. The claims spread rapidly across broadcast and social media. No verifiable evidence of outcome-altering manipulation is produced.
- 2020-11The federal Cybersecurity and Infrastructure Security Agency (CISA), together with the Election Infrastructure coordinating councils, issues a joint statement calling the election the most secure in American history and stating there is no evidence that any voting system deleted or lost votes, changed votes, or was compromised.
- 2020-11Georgia conducts a full statewide hand count of roughly five million paper ballots in the presidential race as a risk-limiting audit. The hand tally of the paper records confirms the machine-counted outcome, with differences well within ordinary human-counting error. Recounts and audits in other states likewise match their machine tallies.
- 2022-05CISA publishes an advisory on vulnerabilities in one line of Dominion equipment found by a security researcher, and states plainly that it has no evidence these vulnerabilities have ever been exploited in any election. The episode captures the whole distinction: a real potential flaw, and no real-world exploitation.
- 2023-04Dominion Voting Systems and Fox News settle Dominion's defamation suit for 787.5 million dollars, after a judge had ruled the network's specific claims about Dominion were false. The settlement and related litigation establish that prominent machine-rigging accusations were made without supporting evidence.
- 2024-2025Around the 2024 election, claims that machines are switching votes go viral again and are rebutted by election officials, security experts, and fact-checkers, who note that isolated reports trace to user error rather than manipulation. Newsmax acknowledges the falsity of certain Smartmatic claims and settles; Smartmatic's suit against Fox continues.
- 2026Heading into the midterm elections, anticipatory versions of the claim circulate in advance, asserting that machines will be used to steal the coming vote. As in prior cycles, the assertions arrive before any results and rest on the general distrust of the technology rather than on evidence of manipulation.
From the case file
The actual records: declassified, released, or leaked. We link straight to each document in its official archive, so you never have to take our word for it. Read the originals yourself.
Contradicted. Two different things are often folded together here. The documented record is that electronic voting machines, like any computer, have measurable vulnerabilities, and that security researchers have demonstrated some of them in controlled lab conditions; this is real and it is why paper ballots and audits exist. The rated claim is far larger: that machines in actual US elections were hacked or pre-programmed to secretly flip or fabricate votes and change who won. That claim is debunked. Post-election audits, hand recounts, and paper-ballot checks have repeatedly matched machine tallies (Georgia's 2020 full hand count is a clear example); federal authorities (CISA and the EAC) and bipartisan state officials found no evidence of machine manipulation altering any presidential result; and major defamation cases established that specific machine-rigging accusations were false and made without evidence. This file accuses no company or person of rigging anything; on the contrary, the point is that the rigging did not happen.
Sources
- 1.Election Security: Rumor vs. Reality, Cybersecurity and Infrastructure Security Agency (CISA)
- 2.Joint Statement from Elections Infrastructure Government Coordinating Council & the Election Infrastructure Sector Coordinating Executive Committees, Cybersecurity and Infrastructure Security Agency (CISA) (2020)
- 3.U.S. finds no evidence flaws in Dominion voting machines were ever exploited, CBS News (2022)
- 4.Historic First Statewide Audit of Paper Ballots Upholds Result of Presidential Race, Georgia Secretary of State (2020)
- 5.Biden Wins Georgia Per AP After State Ends Hand Audit Confirming His Lead, NPR (2020)
- 6.Dominion Voting Systems v. Fox News Network, Wikipedia
- 7.Fox News settles blockbuster defamation lawsuit with Dominion Voting Systems, NPR (2023)
- 8.Smartmatic Settles With Newsmax: Here's Where It And Dominion's Other Lawsuits Stand Now, Forbes (2024)
- 9.False claims about machines "switching" votes are going viral. Here's what to know., CBS News (2024)
- 10.Analysis: How PolitiFact fact-checked 2024 election claims on vote counting, mail ballots, machines, PolitiFact (2024)
Help us investigate
This is a living case file. If you spot an error or know evidence we missed, tell us, and weigh in on where you land.
Where do you land?
Cast your read on this one.