Voter Registration and List Maintenance

In August of 2026, the Georgia Department of Audits and Accounts released a report on Georgia’s voter rolls. The conclusion is not good.

A link to the full report is at the bottom of this article, as are some other very interesting links. But here are the top 9 problems that the report found.

1. Potential noncitizens were placed into active voter status

This is probably the report’s most consequential finding.

The auditors identified approximately 2,600 registrations that had been active or inactive even though citizenship could not be verified through the voter system, DDS records, or federal database searches:

• 1,395 had already been identified through SOS or county reviews.
• The auditors found another 1,175 active or inactive registrations for which the latest DDS record indicated non-U.S. citizenship and the federal SAVE search did not establish citizenship.
• Of those 1,175, 182 had some voting history.
• After further review, 944 were placed in pending status or canceled.
• At the time of the report, 231 remained active or inactive while additional investigation was pending.

Across the larger group of approximately 2,570 potential noncitizen registrations, 407 had a voting history. But that does not establish that 407 noncitizens voted illegally: the report says the recorded votes ranged from 1996 to 2025, and the individual could have been a citizen when voting or the DDS citizenship indicator could have been wrong. (pp. 20–22)

2. DDS erroneously classified some noncitizens as citizens

The underlying control failure was particularly troubling.

DDS employees sometimes:

• Misclassified immigration or employment-authorization documents as birth certificates.
• Failed to obtain necessary citizenship documentation.
• Marked applicants as citizens before REAL ID review was completed.
• Caused an automatic voter-registration application to be submitted based on the erroneous citizenship designation.

The audit examined 359 automatic registrations generated even though the individual’s latest DDS documentation indicated noncitizenship. Fifty-two of these registrations had a voting history.

DDS then reviewed approximately nine million active credential records and identified 3,285 additional REAL IDs that might have been issued without sufficient identity or citizenship documentation. DDS reported that:

• 750 people were verified as U.S. citizens.
• 678 were verified as lawfully present but not citizens.
• 1,857 credentials were canceled because DDS could not verify the necessary information.

An especially important weakness was that when DDS later corrected someone’s status from citizen to noncitizen, there had been no mechanism automatically notifying SOS to review the voter registration. DDS and SOS reported that reconciliation procedures and additional controls were being established. (pp. 22–25)

3. Thousands of voters were registered at addresses that may not establish residency

The audit found several categories of questionable residential addresses:

The first three categories plus the placeholder-address records amounted to nearly 6,800 records that the auditors considered potentially insufficient to establish residency. That is a small fraction of eight million registrations, but an incorrect residence can place a voter in the wrong congressional, legislative, school-board, municipal, or local district. (pp. 14–19)

The report also found approximately 70,000 registrations with potentially defective mailing addresses, although the auditors could not determine how many were truly undeliverable.

4. Duplicate active registrations remained in the system

The auditors identified 1,210 people with more than one registration in active or inactive status:

• 955 had both records designated active.
• Three people appeared to have used both duplicate records to vote in the same election.

Georgia runs nightly duplicate checks, but the audit identified a specific control gap: if a county registrar reviews a potential duplicate pair and erroneously clears it as not being a duplicate, the system will not present that pair for review again.

The three same-election cases are among the report’s clearest indications of potentially improper voting, although the report does not provide enough information to determine whether they involved intentional double voting, data-linking errors, or another explanation. (p. 32)

5. Some registered voters’ identities could not be confirmed

Of approximately 7.89 million registrations containing a REAL ID number, the auditors confirmed 99.8%. Approximately 17,855 could not initially be matched through their automated criteria.

Most of those had another explanation—such as a Social Security number, a minor name discrepancy, or a status requiring additional documentation. Nevertheless, the auditors estimated that between 786 and 1,163 registrations:

• Could not be verified through the listed REAL ID or Social Security information; and
• Were not coded to require additional identity documentation before voting.

The report estimated that no more than 281 people within one projected subgroup likely had voting history. Again, this is not a finding of fraudulent voting: counties may have examined documentation that was not contained in the data supplied to the auditors. (pp. 11–14)

6. Approximately 20,000 records contained illogical “last contact” dates

The audit found about 20,000 records—0.25% of the voter roll—with dates of last contact that were logically impossible or inconsistent, including:

• Dates preceding a more recent vote.
• Dates preceding registration.
• Future dates.

These errors matter because the “last contact” date helps determine whether a voter should receive a no-contact notice and eventually become inactive. An erroneously old date could cause a recently active voter to be treated as having no contact; an erroneous future date could prevent appropriate list maintenance.

The auditors noted that additional dates are reviewed before final cancellation, which reduces the risk of an improper removal. (pp. 31–32)

7. Quality control varies considerably among Georgia’s 159 counties

Although all surveyed counties reported some quality-control activity, the rigor and timing varied:

• Only 55% reported that a second employee compares records against source documents.
• Only 41% reported using GARViS audit reports promptly—daily or weekly.
• Eleven counties reported performing none of the transaction-review procedures listed by the auditors.
• Only 70% reported using GIS comparisons to validate addresses.
• Only 22% reported using the “Not Eligible Georgian” GARViS feature.
• Six metro Atlanta counties reported more than 50,000 voters for each employee assigned to daily registration and maintenance work.

This decentralization means that the same questionable record may be handled differently depending on the county’s staffing, experience, procedures, and interpretation. (pp. 32–36)

8. Georgia lacks comprehensive, current written procedures

The auditors found:

• SOS did not maintain comprehensive formal list-maintenance policies and procedures.
• The principal GARViS manual was outdated.
• Training slides were scattered and did not fully address topics such as felony information received outside SOS channels or detailed quality-control methods.
• Only 27 of 159 counties reported having developed internal written guidance.
• Only 16 counties had standard operating procedures.

SOS disputed the need for statewide formal procedures, arguing that counties are legally responsible and must adapt their procedures to local size and resources. The auditors concluded that clearer statewide guidance would improve consistency. (pp. 35–38)

9. Court reporting of ineligible individuals was inconsistent

Superior and probate courts must report people who:

• Identify themselves as noncitizens during jury qualification; or
• Have been declared mentally incompetent for voting purposes.

SOS reported regular submissions from 35 superior courts and 121 probate courts. The audit could not tell whether the other courts had nothing to report, sent information directly to counties, or simply failed to report. Interviews revealed limited awareness and inconsistent practices.

SOS introduced reporting instructions and a form in November 2025, but the auditors recommended continuing to identify courts with reporting lapses. (pp. 27–30)

To just download the pdf of the report, click HERE

Click HERE to visit the Georgia Department of Audits and Accounts webpage that has this report

Click HERE to see a very well done ‘dashboard’ with LOTS of information