SERVANTSTACK // INCIDENT INTELLIGENCEFORENSIC IMAGE // VERIFIED FRAME

SS-IR-004 // INCIDENT REPORTAlleged

Michigan MiDAS

Automated System Falsely Accuses 40,000 of Unemployment Fraud

Q: Which governance control was missing?

Named SME review and decision audit trail. A qualified reviewer tests the basis, context, and disparate impact before the decision reaches a person.

EXECUTIVE BRIEF

BIAS & ACCESS GOVERNMENT & PUBLIC SECTOR

FAILURE CHAINTRACE COMPLETE

01TRIGGERMichigan's Unemployment Insurance Agency deployed MiDAS (Michigan Integrated Data Automated System), an automated…
02MACHINE ACTIONMaterial contributor
03MISSING GATENamed SME review and decision audit trail
04IMPACTFinancial harm

01 // INCIDENT SUMMARY

The short version

02 // KEY FACTS

Case telemetry

INCIDENT: SS-IR-004
DATE: October 2013 - August 2015
SYSTEM: Michigan MiDAS
LOCATION / SCOPE: Michigan, USA
EVIDENCE: Alleged
AI ROLE: Material contributor
HARM: Financial harm
SOURCES: 1 cited record

03ENTRY POINT // WHAT HAPPENED

The event

Michigan's Unemployment Insurance Agency deployed MiDAS (Michigan Integrated Data Automated System), an automated fraud detection system that cross-referenced employer and claimant data to flag discrepancies. The system generated fraud determinations with a 93% false positive rate. Over 40,000 people were falsely accused of unemployment fraud. MiDAS automatically assessed quadruple penalties - the standard benefit repayment plus a 400% fine - and garnished wages, seized tax refunds, and destroyed credit scores without any human review.

04CAUSAL TRACE // AI'S ACTUAL ROLE

What the machine did

MiDAS operated for 22 months with zero human review of fraud determinations. The system flagged data entry errors, employer reporting mistakes, and legitimate misunderstandings as intentional fraud. It then auto-generated penalty assessments and collection actions. The Michigan UIA had laid off most of its human adjudicators and relied entirely on MiDAS output.

Material contributorAutomation was a causal participant—not a decorative label for the system around it.

05BLAST RADIUS // CONSEQUENCES

Where the failure landed

40,000+ people falsely accused of fraud. $117 million in wrongful penalty assessments. Families lost homes to garnishment. Credit scores were destroyed. Multiple people reported suicidal ideation. The Michigan Auditor General found a 93% error rate - the system was wrong in more than 9 out of 10 cases. A class action settlement followed.

06 // EVIDENCE STATUS

Alleged

Claims reported in litigation or public allegations; not presented here as a final finding.

SOURCE RECORD UPDATED 2026-07-09

07 // SOURCE LEDGER

1 cited record

01
Primary / officialMichigan Office of the Auditor General: UIA Fraud Investigation (2017)

08CONTROL FAILURE // MISSING GOVERNANCE

Named SME review and decision audit trail

The failure pattern in this case: Automated judgment without accountable review.

09INTERVENTION POINT // HUMAN IN THE MIDDLE

The moment the path could change

A qualified reviewer tests the basis, context, and disparate impact before the decision reaches a person.

AI PROPOSES→HUMAN OWNS THE DECISION→SYSTEM EXECUTES

10CONTROL DEPLOYMENT // AUTHORITYGATE

SME routing · decision audit trail

AuthorityGate's framework requires human adjudicator review before any automated fraud determination triggers financial penalties. The original system had human adjudicators - MiDAS replaced them. The framework treats fraud accusations as high-stakes decisions that require SME validation. A human reviewer would have seen that a claimant's employer reported slightly different hours - a data entry issue, not fraud.

RELEVANT KEYSTONE CONTROLHuman-in-the-Loop ValidationHow high-risk actions route to a named subject-matter expert who owns the go or no-go decision.

12 // THE ALTERNATIVE

Autonomy is a design choice.

See the operating model that keeps AI useful while preserving human authority at consequential moments.

Compare AgenticAI and AugmentedAI →