Checkr: $4M FCRA misattributed record resolution

Checkr Inc., one of the largest automated employment background screening companies in the United States and a primary screening provider for major gig economy platforms, agreed to pay $4 million to resolve Fair Credit Reporting Act charges that its algorithmic record-matching procedures attributed criminal, civil, and other adverse records from one individual to a different person being screened — generating consumer reports that harmed job applicants who had no such history in their own records.DOCUMENTED

Checkr's technology-driven model processes high volumes of reports by comparing name, date of birth, Social Security number, and address history against criminal and civil databases and using match scores to determine record attribution. Even a small algorithmic error rate across millions of reports can affect large numbers of individuals — and for the affected applicant, the impact of receiving someone else's criminal record in their background report is complete, not statistical.

How Automated Matching Produces Errors

Checkr's matching logic weights combinations of identifying data points to determine whether a database record belongs to the individual being screened. Misattribution occurs most frequently when two people share similar names and one or more overlapping data points — same birth year, prior address in the same county, partial Social Security number match. A criminal court record entered under a common name with limited identifying information may match multiple individuals' profiles, and without additional verification steps, the algorithm may attribute it to the wrong person.REVIEWED

Regulators found that Checkr's procedures did not include sufficient human review or supplemental verification when match confidence was ambiguous. The FCRA's requirement to follow reasonable procedures to assure maximum possible accuracy was violated in the category of cases where incomplete or overlapping identifying information made incorrect attribution foreseeable. At Checkr's processing volume, even rare error types affect substantial numbers of applicants in absolute terms.DOCUMENTED

The Gig Economy Amplification Effect

Checkr's prominence among gig economy platforms meant errors compounded across platforms. An applicant denied by one platform due to a misattributed record who applied to a second platform using Checkr faced the same inaccurate report and the same denial — because both platforms independently drew on the same algorithmic matching process. An error that affects one report affects all simultaneous applications to Checkr-using platforms, multiplying the employment impact of a single algorithmic mistake across a pool of platforms on which the applicant may have depended for income.REVIEWED

The algorithm produced the same error for every platform that ran a check on the same applicant. Each denial looked independent. The underlying cause was identical.

Settlement Reforms

The settlement required Checkr to implement enhanced verification steps for cases where match confidence is ambiguous — adjusting thresholds for automated attribution when database records lack key disambiguating identifiers. Human review requirements were added for borderline cases. Dispute resolution timelines for misattributed-record claims were shortened, and procedures were established to notify employers who received the original inaccurate report when a corrected version is produced, giving them an opportunity to reconsider employment decisions made on inaccurate information.DOCUMENTED

Applicant Rights Under the FCRA

Applicants who receive an adverse employment decision based on a background report are entitled to a pre-adverse action notice with a copy of the report before the decision is finalized. This notice creates an opportunity to dispute inaccurate information. Applicants who identify records that belong to a different person can dispute directly with Checkr, which must investigate within 30 days and correct or delete unverified attributions. Consumer rights attorneys who specialize in FCRA claims can evaluate whether procedural violations in the adverse action process support a claim for statutory damages up to $1,000 per violation plus attorneys' fees.

For employers who rely on automated screening services, the Checkr case reinforces that they cannot treat background reports as infallible determinations. When an applicant disputes accuracy — particularly when they dispute records that appear to belong to a different individual with a similar name — the employer has both a legal obligation and a practical interest in allowing the dispute process to complete before making a final adverse action decision. Employers who receive a corrected report after an adverse action has been taken should have a process for revisiting that decision. Auditing vendor relationships for FCRA compliance — asking for documentation of accuracy procedures, dispute resolution processes, and error rate monitoring — is reasonable due diligence for any employer processing significant background check volume.

The settlement marks an important precedent for how the FCRA's accuracy requirements apply to algorithmic consumer reporting systems at scale. The statutory obligation to maintain reasonable procedures for maximum possible accuracy does not become less demanding because the volume of reports makes individual human review impracticable. Instead, the reasonableness standard requires that algorithmic systems incorporate the quality-control disciplines — thresholds, exception queues, supplemental verification — that a human reviewer would apply when confronted with an ambiguous match. The settlement's requirements establish a compliance standard for the automated screening industry that will inform future regulatory expectations as the technology and the market continue to evolve.

For employers who rely on automated screening services, the Checkr case reinforces that they cannot treat background reports as infallible determinations. When an applicant disputes accuracy — particularly when they dispute records that appear to belong to a different individual with a similar name — the employer has both a legal obligation and a practical interest in allowing the dispute process to complete before making a final adverse action decision. Employers who receive a corrected report after an adverse action has been taken should have a process for revisiting that decision. Auditing vendor relationships for FCRA compliance — asking for documentation of accuracy procedures, dispute resolution processes, and error rate monitoring — is reasonable due diligence for any employer processing significant background check volume.

The settlement marks an important precedent for how the FCRA's accuracy requirements apply to algorithmic consumer reporting systems at scale. The statutory obligation to maintain reasonable procedures for maximum possible accuracy does not become less demanding because the volume of reports makes individual human review impracticable. Instead, the reasonableness standard requires that algorithmic systems incorporate the quality-control disciplines — thresholds, exception queues, supplemental verification — that a human reviewer would apply when confronted with an ambiguous match. The settlement's requirements establish a compliance standard for the automated screening industry that will inform future regulatory expectations as the technology and the market continue to evolve.

Have documents relevant to this story? Reach us through our tips channel.