Confirmed, Debunked, Unresolved: The 2020 Election Irregularities the Media Stopped Covering

The 2020 presidential election outcome has been confirmed by every recount, audit, court, and law enforcement. That is the factual baseline and it is not in dispute. What is in dispute — and what has been almost entirely absent from national media coverage — is whether the process was clean. It wasn't. Across Georgia, Arizona, Pennsylvania, Michigan, and California, state election boards, federal investigators, and official audits documented real procedural failures: broken chains of custody, data entry errors in audits, last-minute court-ordered rule changes, confirmed breaches of voting equipment, and in one case a state political party operating fake ballot drop boxes. These failures did not change the 2020 outcome. They do represent vulnerabilities that have not been adequately fixed. This piece documents what actually happened — not what partisans on either side claimed happened — so voters know what to watch for next time.

Georgia: The Most Documented Case

Georgia conducted three separate counts of its nearly five million ballots — the original machine count, a full statewide hand recount, and a machine recount requested by the Trump campaign. All three confirmed Biden's margin of victory. Georgia's Republican governor, Republican Secretary of State, and Republican Attorney General all certified the result.

The administrative record in Fulton County, Georgia's largest county, is a separate matter. It is documented, official, and largely unreported.

Unsigned tabulator tapes covering 315,000 votes | CONFIRMED

In December 2025, Fulton County acknowledged that poll workers failed to sign approximately 134 tabulator tapes covering roughly 315,000 early votes. Georgia rules require these signatures as certification that each tape accurately reflects a machine's count — they are the paper trail that makes a machine recount independently verifiable. The State Election Board referred the matter to the Attorney General and recommended fines exceeding $670,000, at $5,000 per violation.

The Secretary of State called it "sloppy administration" and maintained the votes remain valid because ballots had been verified through voter ID and signature matching before counting. Critics argue that without the signed tapes, independent certification of those 315,000 votes is impossible after the fact. Both positions have merit. What is harder to defend is that this took years to surface and received almost no national attention.

Double-scanning of ballots during recount | CONFIRMED

During the Trump-requested machine recount, approximately 3,000 ballot images were scanned twice due to poor batch management. An FBI search warrant affidavit in January 2026 cited this as a potential "defect." The State Election Board issued a formal letter of reprimand to Fulton County for violating batch management procedures. Investigators concluded the error did not affect the certified totals, which were confirmed by the hand audit. The confirmed finding is the procedural violation — a documented breach of the rules governing the recount, not evidence of deliberate manipulation.

Data entry errors in the hand-count audit | CONFIRMED

During the statewide risk-limiting audit designed to check the machines, Fulton County staff duplicated batch entries and misidentified scanner batches when uploading totals to the state's auditing software. The State Election Board found the county in violation of audit rules and reached a consent order in 2023 requiring new training. The body meant to verify the machines had its own documented errors. Officials noted the hand count is a statistical check rather than a precise recount and that the certified result relied on the machine recount. That is technically accurate and editorially insufficient — an audit with its own data entry errors is an audit whose findings are harder to defend.

Missing ballot images and authentication files | CONFIRMED as recordkeeping failure

Investigations found that ballot images for some absentee ballots were missing from the public record, and SHA authentication files — digital signatures used to verify the integrity of ballot images — were not properly generated or preserved. The FBI's 2026 seizure of Fulton County election equipment targeted these specific records. State investigators attributed the gaps to server migration issues and recordkeeping failures rather than deliberate deletion. The physical paper ballots were intact and available. The digital record, which exists precisely because paper is not always sufficient, was not.

The "suitcase ballots" allegation | DEBUNKED

Video from State Farm Arena appeared to show election workers pulling "suitcases" of hidden ballots from under a table after observers left, then scanning them repeatedly. The claim spread to millions of viewers and resulted in documented harassment, death threats, and years of persecution directed at workers Ruby Freeman and Shaye Moss.

A joint investigation by the Georgia Secretary of State's office, the Georgia Bureau of Investigation, and the FBI found the allegation entirely false. The containers were standard ballot transport cases. The workers were following standard procedure. The State Election Board dismissed the case. Freeman and Moss later won significant defamation settlements. This is the clearest example in the entire 2020 record of a viral false claim causing direct harm to private citizens who did nothing wrong.

The Coffee County equipment breach | CONFIRMED

This story received far less attention than the debunked suitcase allegation. Two months after the 2020 election, in January 2021, a team of outside computer experts traveled to rural Coffee County and spent hours inside the local elections office copying software and data from Dominion voting equipment — servers, voter sign-in devices, memory cards, and election records. Security cameras captured their arrival. Local election officials were present during the access.

The person who facilitated their entry was Cathy Latham, then chair of the Coffee County Republican Party — the same individual simultaneously participating in Georgia's fake elector scheme, for which she was later indicted on racketeering charges. Security footage also showed Cyber Ninjas CEO Doug Logan, later hired to audit Maricopa County in Arizona, visiting the office weeks afterward. The trip was reportedly directed by former Trump attorney Sidney Powell.

When the nonprofit Coalition for Good Governance sued Coffee County over the breach, county officials withheld emails and security footage for months, at one point denying the records existed before turning them over to the Georgia Bureau of Investigation after a direct demand. A federal court sanctioned the officials for withholding evidence. The Georgia Secretary of State's office later acknowledged it had failed to properly detect or investigate the breach.

Outside parties gained unauthorized physical access to certified election equipment. They copied its contents. Local officials helped it happen and then concealed evidence of it. The story received almost no national coverage during the same weeks that the debunked suitcase claim was being discussed on every major network.

Broader administrative failures | CONFIRMED

The Performance Review Board identified a systemic pattern of poor records management across Fulton County's 2020 operation: ballots sent in individual batches rather than complete boxes during the recount, leading to temporarily lost batches; inadequate staffing for the highest-turnout election in Georgia history; and time pressure that allowed errors to compound undetected. None of this was attributed to intentional manipulation. The board's own framing — organizational failure under extraordinary circumstances — raises a question that went largely unasked: is "extraordinary circumstances" an acceptable explanation for the county administering a contested presidential election?

---

Arizona: The Audit That Confirmed What It Was Supposed to Overturn

The Cyber Ninjas audit | CONFIRMED as confirming Biden's win

Arizona Senate Republicans hired Cyber Ninjas — a Florida cybersecurity company whose CEO had publicly promoted Stop the Steal content — to conduct a comprehensive review of 2.1 million Maricopa County ballots. The audit became a months-long spectacle. Counters initially used blue pens, which are prohibited in audits because the ink can be read by tabulators; a court halted the process on its first day. When the audit concluded, Cyber Ninjas found Biden had won — by a wider margin than the official count, giving Biden 99 additional votes and Trump 261 fewer than the certified results.

Maricopa County issued a 93-page rebuttal documenting specific methodology errors in the Cyber Ninjas work, including "soft matches" — flagging two entirely different voters as potential double-voters because they shared the same first name, last name, and birth year. The county had already been audited by two federally certified firms before Cyber Ninjas arrived. The state ultimately spent millions in taxpayer funds replacing election equipment that had to be scrapped after being handled by outside contractors without standard security protocols.

Arizona's Republican Attorney General suppressed fraud findings | CONFIRMED

This is one of the least-reported confirmed facts of the entire 2020 story. Arizona's Republican Attorney General Mark Brnovich commissioned a thorough investigation of fraud allegations — more than 10,000 investigator hours, 638 complaints reviewed, 430 investigations opened, 22 cases referred for prosecution. In March 2022, investigators concluded in an internal memo that most fraud claims were "speculative" and "found to be inaccurate" when investigated. Brnovich never released the memo. A month later, he released a public "interim report" suggesting his investigation had "raised questions" about the election.

His Democratic successor, Kris Mayes, released the suppressed documents in early 2023. The record is clear: a Republican attorney general, facing a primary electorate that demanded fraud findings, sat on 10,000 hours of investigative work that cleared the election, issued misleading public statements implying the opposite, and left office without disclosure. That story received a fraction of the coverage given to the original fraud allegations it was designed to investigate.

Fake electors | CONFIRMED as documented scheme

Arizona was one of seven states where Trump allies signed false Electoral College certificates claiming Trump had won states he had lost. Eighteen people were charged in Arizona, including 11 fake electors and former Trump aides Rudy Giuliani and Mark Meadows. Trump was listed as an unindicted co-conspirator. This is not ballot fraud — it is a documented post-election attempt to substitute fraudulent electoral certificates for legitimate ones, a separate and serious form of election interference that courts in multiple states have addressed.

---

Pennsylvania: Rules Changed Mid-Election

Pennsylvania's 2020 election was the most legally chaotic of any battleground state. The source of that chaos was not fraud by voters or workers — it was a contested judicial rewrite of the rules weeks before Election Day.

Extended mail ballot deadline | CONFIRMED as court-ordered rule change

In September 2020, the Pennsylvania Supreme Court, in a 4-3 decision split along party lines, extended the deadline for receiving mail ballots by three days — to November 6, three days after Election Day — and ruled that ballots without postmarks would be presumed valid. The legislature had not authorized this change. The U.S. Supreme Court declined to block it before the election but noted the legal question was "weighty." The extension affected roughly 10,000 ballots, too few to change Pennsylvania's outcome. The mechanism — a court rewriting election rules weeks before a major election — is a legitimate structural concern regardless of which party it benefited.

Signature verification suspended | CONFIRMED as deliberate policy decision

Pennsylvania's Department of State issued guidance telling counties they could not reject mail ballots based solely on signature comparison. The Pennsylvania Supreme Court affirmed that position in October 2020. One of the traditional verification mechanisms for mail voting was deliberately suspended for the highest-turnout election in the state's history. This was public policy, not a covert act. But no public accounting was ever produced of how many ballots would have been flagged under prior procedures.

Data irregularities in the state voter database | UNRESOLVED

Pennsylvania state Rep. Frank Ryan, a Republican CPA, testified to the U.S. Senate in December 2020 that he had identified approximately 202,000 more votes in the state's SURE voter database than voters recorded as having cast ballots. He also identified 154,584 ballots with mail date irregularities — including some dated as mailed before the ballot itself existed — and 1,573 records showing voters over age 100. Ryan called for access to the SURE system's transaction logs to determine whether the discrepancies were data entry errors, system artifacts, or something else.

The Pennsylvania Department of State attributed the discrepancies to database timing issues and record-keeping delays. A full public reconciliation was never conducted to the satisfaction of independent observers. The underlying data was never made fully public. Whether these numbers represent a legitimate unresolved question or a thoroughly explained technical footnote depends on records that were never transparently shared.

---

Michigan: The Window Covering That Wasn't What It Looked Like

Detroit TCF Center | MISLEADING FRAMING

Images and video of election workers covering the windows of the TCF Center counting room in Detroit circulated widely on election night and became one of the most-referenced visuals in the fraud debate. The optics, stripped of context, were genuinely suspicious.

The full context is documented. Outside the building, a crowd had gathered in response to social media calls to converge on the count — people were banging on the windows and pointing phones at workers inside, photographing and recording them at close range through the glass. Michigan law prohibits unauthorized recording of election workers and voters in ways that expose private information. The city's legal counsel ordered partial window coverings in response to workers' requests to stop being filmed. Not all windows were covered. Dozens of credentialed observers from both parties, plus reporters, were inside the room throughout the process and observed everything.

Wayne County Circuit Chief Judge Timothy Kenny, who heard the lawsuit that followed, found that challengers claiming denial of access "did not have a full understanding" of the vote counting process and that their account of events was "incorrect and not credible." Detroit officials later acknowledged the room's layout — which had exterior-facing windows because the space was improvised due to COVID — was a design mistake. For subsequent elections, the count was moved to an interior space specifically to prevent a repeat of those conditions.

The covered windows were not fraud. They were a foreseeable consequence of placing a high-stakes ballot count in a room visible from outside, during an election in which the losing campaign had spent months priming its supporters to treat any unexplained event as evidence of theft.

Fake electors | CONFIRMED, charges later dismissed on procedural grounds

Michigan charged 16 fake electors, including a former co-chair of the state Republican Party, with forgery and election law violations. A federal judge dismissed the charges in September 2024 on jurisdictional grounds, not on the merits. The underlying conduct — signing and submitting fraudulent electoral certificates — is documented.

---

California: Three Stories the State Didn't Want to Tell

California was not a contested state in 2020 — Biden won by more than five million votes — and received no scrutiny in the fraud debate. That does not mean nothing happened. Three separate California stories from 2020 illustrate that election irregularities are not a partisan phenomenon.

The California Republican Party's fake drop boxes | CONFIRMED

In October 2020, the California Republican Party acknowledged placing more than 50 unauthorized ballot drop boxes across Fresno, Los Angeles, and Orange counties, labeled "Official Ballot Drop-off." California law limits the authority to establish official drop boxes to county election officials exclusively. The Secretary of State called the boxes illegal, issued a cease-and-desist order, and the Attorney General opened a formal investigation. When the California Republican Party refused to comply with investigative subpoenas, the AG sought a court order compelling compliance.

The party agreed to remove the "official" labeling. No criminal charges were filed. Courts later ruled that unofficial party-run collection was not per se illegal without the misleading designation.

The factual record: a state political party that spent 2020 warning the public about drop-box fraud operated more than 50 unauthorized drop boxes designed to look like official government equipment, deceiving voters about where their ballots were going. The investigation was quietly dropped after the party agreed to change a label. The story briefly made headlines in October 2020 and then largely disappeared.

The Hawthorne gang voter registration scheme | CONFIRMED, prosecuted

In November 2020, the Los Angeles County District Attorney charged Carlos Antonio De Bourbon Montenegro and co-defendant Marcos Raul Arevalo with 41 combined criminal counts. Montenegro, running for mayor of Hawthorne under a front nonprofit called "Homeless Votes Matter," allegedly submitted more than 8,000 fraudulent voter registration applications between July and October 2020, listing fictitious, deceased, and nonexistent persons as voters, using three post office boxes and his own home address. Prosecutors alleged the scheme was funded by the criminal gang MS-13, which sought to install a sympathetic candidate in local office.

The scheme was detected by the LA County Registrar-Recorder, which flagged the suspicious volume of registrations at common addresses. Montenegro pleaded no contest to three counts and was sentenced to 60 days in jail, two years' probation, and community service.

This case is important for what it demonstrates about how fraud detection works when it works. The volume triggered automated flags. Investigators traced it. Prosecution followed. Montenegro's 8,000 applications did not translate into 8,000 votes. The question worth asking is whether those detection systems are consistently staffed and resourced — particularly in a state that automatically registers voters through DMV transactions.

California DMV automatic registration and noncitizen registrations | CONFIRMED as systemic failure

In 2018, the California DMV's automatic voter registration system generated more than 100,000 erroneous registration records — wrong party assignments, incorrect mail ballot designations, and approximately 1,500 registrations sent to people ineligible to vote, including noncitizens. A subsequent state review confirmed that six of those ineligible registrants had voted in the 2018 midterm elections before their registrations were identified and canceled.

The DMV attributed the errors to agents failing to correctly close registration software between customer appointments, causing data fields to auto-populate with the wrong person's information. Software fixes and additional training were ordered. The ineligible registrations involved legal permanent residents, not undocumented immigrants.

The relevance to 2020 is this: the same motor-voter infrastructure, patched but architecturally unchanged, was in place when California ran its 2020 election. The state has never produced a comprehensive public reconciliation confirming that no residual ineligible registrants from the 2018 errors remained on rolls when 2020 ballots were distributed. California Secretary of State Alex Padilla, now a U.S. Senator, called the original errors "absolutely unacceptable" and said they risked "jeopardizing confidence in the electoral process." That statement was accurate in 2018. It remains accurate.

---

The Dominion Voting Machines Question

This subject requires precise distinctions. The way it was covered — first dismissively, then partially, then selectively — made honest public understanding of it nearly impossible.

What the conspiracy theory claimed, and why it failed

After the 2020 election, claims spread that Dominion Voting Systems machines had been programmed to flip votes, connected to the internet during counting, or remotely manipulated by foreign actors. CISA, state auditors, independent researchers, and courts all investigated. None found evidence of vote-switching, remote access, or exploitation of machines during the 2020 election. Every recount and audit confirmed machine tallies against paper records. Fox News paid $787.5 million to settle Dominion's defamation lawsuit, with internal communications showing executives and on-air personalities knew the claims were false while continuing to broadcast them. That is the record on those specific claims.

What is also true: these machines have known, documented vulnerabilities

In 2022, CISA confirmed nine specific software vulnerabilities in certain Dominion ImageCast X ballot-marking devices used across 16 states. University of Michigan professor J. Alex Halderman demonstrated in a sealed Georgia federal court proceeding that the flaws were genuine and could allow a malicious actor to alter the bar codes machines read when tabulating votes.

Since 2017, the annual DEF CON security conference has hosted a "Voting Village" where experienced security researchers are given access to voting machines used in real elections. The findings across every year have been consistent: with physical access and sufficient time, these machines can be compromised. Vote tallies can be altered. Unauthorized software can be installed. The 2017 report concluded bluntly that "even with limited resources, time, and information, voting systems can be hacked." Members of Congress from both parties have attended and observed these demonstrations firsthand.

This is not a fringe argument. In 2018, then-Senator Kamala Harris stated at a Senate Judiciary Committee hearing: "I actually held a demonstration for my colleagues here at the Capitol where we brought in folks who, before our eyes, hacked election machines."

Security researchers and election vendors offer a legitimate counterpoint: DEF CON's conditions — hotel ballroom access, hours of uninterrupted time, no election-day security protocols in place — don't replicate the real-world environment. Layers of physical security, pre-election logic and accuracy testing, post-election paper ballot audits, and chain-of-custody controls all reduce real-world risk substantially. CISA found no evidence the documented vulnerabilities were exploited in 2020.

The honest conclusion is that certified voting equipment used to process tens of millions of votes contains documented vulnerabilities that credentialed researchers have repeatedly demonstrated can be exploited under the right conditions. "Not exploited yet" and "unexploitable" are not the same statement. The policy question — whether the United States should continue relying on software-dependent voting equipment rather than hand-marked paper ballots — is legitimate, bipartisan, and unresolved.

---

What the Courts Actually Found

More than 60 lawsuits challenging the 2020 results were filed in battleground states. The breakdown is worth understanding in specific terms, not just as a round number.

Roughly 14 cases were withdrawn by Trump's legal team before courts ruled. Of those that reached federal judges, 47 were dismissed — including by judges appointed by Republican presidents, Democratic presidents, and Trump himself. At least eight Trump-appointed judges dismissed fraud claims. The Third Circuit panel that rejected the Trump campaign's Pennsylvania case was composed entirely of Republican-appointed judges. A judge in Arizona dismissed a case from the bench during an evidentiary hearing and ordered the Arizona Republican Party to pay legal fees for filing what the court called a "groundless" bad-faith lawsuit.

Of the approximately 30 cases heard on the merits, all but one were dismissed. Courts identified the same problems repeatedly: allegations were too vague to adjudicate — no specific locations, no named actors, no documented frequency of alleged misconduct. Evidence consisted largely of statistical arguments or expert opinion rather than direct proof of manipulation. And claimed irregularities fell far short of the scale needed to affect any state's outcome.

One clarification matters here. Courts dismissing fraud claims is not the same as courts finding elections were perfectly administered. Courts dismissed claims for lack of evidence that outcomes were affected — a higher legal bar than whether processes were properly followed. The administrative failures documented in this piece — in Georgia, Pennsylvania, and elsewhere — were not litigated on their merits in most cases. They emerged later, through state board proceedings, investigative reporting, and federal search warrants. A court finding no fraud sufficient to overturn an election is not a finding that everything went according to procedure.

Trump's team did win cases. Three, all in Pennsylvania: 270 provisional ballots lacking required signatures were set aside, Election Day provisional ballots were separated from those cast afterward, and the deadline for absentee voters to present ID was extended by three days. None affected Pennsylvania's outcome.

---

The Mail Ballot Question

A 2023 Heartland Institute/Rasmussen Reports poll asked respondents who said they voted by mail in 2020 about specific behaviors. Twenty-one percent said they had filled out a ballot in part or full on behalf of a friend or family member. Seventeen percent said they had voted in a state where they were no longer a permanent resident. Seventeen percent said they had signed a ballot envelope on behalf of someone else.

This is not a survey in which people spontaneously confessed to crimes. It is a structured poll asking about specific actions, some of which are illegal, some of which exist in legal gray areas depending on the state, and some of which respondents may have described inaccurately or incompletely. Filling out a ballot for an elderly spouse who requested help is treated differently in different jurisdictions. The poll results do not prove 14 million fraudulent ballots were cast.

What the poll does reflect — and what the unprecedented scale of mail voting in 2020 created conditions for — is that millions of voters engaged in practices ranging from legally ambiguous to clearly impermissible, and the systems in place were not consistently designed to catch them. Whether any of that rises to outcome-altering fraud is a separate question from whether those practices should be better regulated.

The prosecuted case count is much smaller. The Heritage Foundation's election fraud database, tracking cases over multiple decades, totals approximately 1,400 proven cases across all election types combined. The FBI reported 23 election fraud cases involving the 2020 election, with six involving ballot tampering. The gap between survey-reported behavior and documented prosecution is large. Both figures belong in any honest accounting of what is actually known.

---

The Media Coverage Problem

This piece has noted repeatedly that specific stories received little national attention. That pattern deserves to be stated plainly rather than just implied.

The debunked suitcase ballot allegation in Georgia — which led to years of harassment of two private citizens — received weeks of continuous coverage on national networks and generated millions of social media engagements. The Coffee County equipment breach — in which outside parties physically accessed and copied certified voting equipment with the assistance of local officials, who then concealed evidence of it — was covered briefly in 2022 and largely forgotten.

The Arizona Republican attorney general suppressing 10,000 hours of investigative work that cleared the election received a fraction of the coverage given to the original fraud claims those investigators examined. The California Republican Party's unauthorized drop box operation received a few days of coverage in October 2020 and then disappeared.

The asymmetry is not a minor editorial failure. It is the reason public understanding of 2020 election integrity remains deeply distorted in both directions — with a significant portion of the population believing the election was stolen based largely on claims that were debunked, and another significant portion believing no legitimate concerns exist based largely on a failure to cover concerns that were confirmed. Both misunderstandings are worse for having been reinforced by coverage that selected for drama over documentation.

---

What Would Make Elections More Secure

The following reforms are supported by election security experts across the political spectrum and do not require accepting any fraud theory to endorse.

Voter ID. Approximately 176 nations and jurisdictions require some form of identification to vote. In Europe, 46 of 47 nations require photo ID — the United Kingdom joined that group in 2023. Mexico uses a biometric national voter ID card. Canada, most of South America, India, Israel, and South Africa all maintain voter identification requirements. The argument that voter ID is inherently suppressive is difficult to sustain against the reality that virtually every mature democracy on earth has concluded it is compatible with broad voter access. Implementation design matters — whether free IDs are available, whether multiple forms of identification are accepted, whether alternatives exist for those without standard documents. But the principle that a voter should be able to confirm their identity before casting a ballot is one the international community has broadly settled.

Paper ballot requirements for all machine-marked votes, with mandatory post-election audits comparing paper records to digital tallies. In 2020, 98 percent of votes were cast with paper in some form. That percentage should be 100, standardized nationally.

Statutory clarity on mail ballot curing, signature verification, and drop box security, established by legislatures before elections rather than by courts during them. What happened in Pennsylvania — where the rules governing millions of ballots were rewritten by a court weeks before Election Day — should not be the mechanism by which election procedures are determined.

Standardized ballot image preservation and SHA authentication requirements, so digital records are independently verifiable years after an election. The gaps in Fulton County's digital record should be impossible by statute.

Independent, nonpartisan chain-of-custody auditing for counties above a population threshold, conducted by parties external to the offices being audited and funded by the state.

Mandatory voting machine security patching and independent penetration testing on a defined schedule, with public disclosure of known vulnerabilities and documented remediation — the same standards applied to any other critical government infrastructure.

Published, reconciled voter roll data after each election, sufficient for independent researchers to verify counts against registration records. What Rep. Ryan identified in Pennsylvania's SURE database — whether ultimately explainable or not — should never be a question that can only be answered by officials who have every incentive to not answer it.