THE SYNTHETIC OUTLAW INDEX · SECOND ORDER ANALYSIS · WHAT THE HEADLINE NUMBERS HIDE · MARCH 2026

The Official Story
and What Is Actually Happening

AI displacement looks manageable in the aggregate. The headline numbers are real. So is what sits underneath them. These are the second-order effects: the harms that will not appear in any dataset until the window to address them has already closed.

BY JONATHAN GROPPER, JD · FULBRIGHT SPECIALIST IN AI AND GOVERNANCE · THE SYNTHETIC OUTLAW (FORTHCOMING)
The damage from AI displacement is already happening, concentrated on the people least able to see it coming or fight back: the governance systems designed to stop it will not arrive until after it becomes permanent.
THE VISIBILITY GAP · 2025 DISPLACEMENT
CITED IN OFFICIAL DATA 55K
ACTUALLY DISPLACED (EST.) 200–300K
4–6× MEASUREMENT GAP · CHALLENGER (CITED) VS INDEPENDENT DUAL-METHOD ESTIMATE · SILENT DISPLACEMENT IS NEVER ANNOUNCED
FOR POLICYMAKERS, AGENCIES AND LEADERS RESPONSIBLE FOR AI GOVERNANCE

Synthetic Outlaw analysis identifies the most critical emerging institutional risk: AI systems and automated decision processes which can satisfy formal compliance while defeating the purpose of the rules meant to govern them.

The framework is designed for policymakers, agencies, companies, and institutions confronting second-order AI risk, procurement exposure, accountability gaps, and failures of public trust.

REQUEST A BRIEFING OR ADVISORY CONVERSATION →
01

The Timeline of Invisibility

When harm begins, when it becomes visible in data, when the window to act closes. The consistent pattern: by the time we can confirm the damage, the policy window has already passed.
HARM TIMELINE 2022–2032 · EACH ROW = ONE SECOND-ORDER EFFECT · SOLID = HAPPENING NOW
HARM UNDERWAY, NOT YET VISIBLE IN DATA VISIBLE: WINDOW TO ACT WINDOW CLOSED: BECOMING IRREVERSIBLE NOW
02

The Official Story vs Ground Truth

Six numbers that look reassuring at first read. Each one has something sitting underneath it.
03

The Cascade

These are not six independent problems. Each failure feeds the next. The SO nodes are where governance failure locks the damage in permanently.

SO = Synthetic Outlaw. These are the sectors where AI achieves prohibited outcomes through formally compliant means, with no single party accountable for the harm. At these nodes, the feedback loop becomes structurally irreversible: no appeal mechanism, no accountability chain, no governance framework that maps.

04

Where Harm Concentrates

Displacement and governance failure do not distribute randomly. They route toward populations with the least capacity to make the harm visible or contest it.
SO = SYNTHETIC OUTLAW   AI that produces a prohibited or socially destructive outcome while remaining formally compliant or operating beyond the effective reach and control of the laws and institutions meant to constrain it. Governance failure is scored across four dimensions: Bypass, Diffusion, Capture, and Governance Gap. The Index applies them to individual AI deployments; this page estimates exposure at sector level. Explore the full index →

Each sector is plotted by AI disruption severity (x-axis) against estimated political power to respond (y-axis). The bottom-right quadrant: high disruption, low power. This is where Synthetic Outlaw concentration is most acute, and where sector exposure is highest. The number inside each bubble is the sector exposure estimate (0 to 10): how far a whole labour sector sits inside synthetic outlaw conditions. It is a different measure from the Index's SO risk score, which scores individual AI deployments rather than sectors. The crosswalk below shows where the two meet.

SO CRITICAL (7.5+) · HARM CONCENTRATES HERE SO HIGH (7.0–7.5) SO MODERATE · LABOR STORY DOMINATES
BOTTOM-RIGHT: HIGH DISRUPTION, LOW POWER
Criminal justice and benefits (exposure 8.3 to 8.5): Hundreds of thousands of pretrial, sentencing, benefits eligibility, and child welfare decisions annually are shaped by algorithmic tools with no explanation requirement and no clear accountability chain. Workers are not primarily replaced; the people they served are harmed.
Healthcare admin, prior auth (exposure 8.8): nH Predict denied claims at scale. The Senate confirmed in October 2024 that the denial rate was the product's selling point, not a flaw.
Gig economy (exposure 8.3): Algorithmic scheduling built to fall outside labor law. Prop 22 formalized the misclassification. The harm is the designed product architecture, not a side effect.
TOP-RIGHT: HIGH DISRUPTION, HIGH POWER
Legal services and financial analysis: Highest raw disruption scores. These sectors have bar associations, congressional relationships, and the ability to shape the regulatory response to their own displacement.
Software engineering: High disruption, but the workers have platforms to document and amplify the harm. The entry-level collapse is visible precisely because these workers could make it visible. AI roles reflects occupational concentration: administrative, secretarial, and healthcare admin roles are disproportionately female and disproportionately automatable.

Crosswalk to the Index

SECTOR ESTIMATE vs DEPLOYMENT SCORE · 7 OF 14 SECTORS HAVE A COUNTERPART

The two instruments measure different objects. The Index scores an AI deployment operating under a specific rule regime. This page scores a labour sector, which may contain several such deployments alongside work that involves none. Seven of the fourteen sectors here have a direct counterpart in the Index; the remaining seven, including software engineering, customer service and marketing, are occupational categories with no single deployment to score.

SECTOR (THIS PAGE)EXPOSUREINDEX DEPLOYMENTSO RISK
Healthcare admin8.8Prior Authorization AI8.25
Criminal justice8.5Recidivism Prediction7.75
Financial analysis8.5Algorithmic Trading8.25
Gig economy8.3Gig Economy Classification8.25
Benefits eligibility8.3Benefits Eligibility AI8.00
Insurance8.2Insurance Underwriting7.00
Tenant screening7.5Tenant Screening AI7.00

Sector estimates run slightly above their deployment counterparts, which follows from scoring a whole sector rather than one system. Gig economy adopts the Index reading directly.

"

The Synthetic Outlaw governance failure does not distribute harms evenly. It routes them toward people who cannot easily make the harm legible, then dissolves the accountability chain that would let anyone be held responsible. Criminal justice. Benefits eligibility. Prior authorization. Tenant screening. These sectors will not lobby Congress. They will absorb the damage quietly, because the people harmed have no platform from which to amplify it.

The sectors scoring highest on this index share a structure, not a coincidence. Bypass operates through compliant means. Diffusion dissolves accountability across vendors, models, and institutions. Capture ensures the regulated shape the regulation. As these systems improve and embed, that structure does not weaken. It consolidates. What is visible today is not the problem at scale. It is the proof of concept.

JONATHAN GROPPER, JD · THE SYNTHETIC OUTLAW (FORTHCOMING)
SOURCES

BLS Employment Projections (Jan 2026) · Challenger Gray and Christmas (full year 2025) · ITIF (Dec 2025) · PwC AI Jobs Barometer (2025) · Goldman Sachs (Aug 2025) · Yale Budget Lab (Feb 2026) · Stanford Digital Economy Study · Hui, Reshef & Zhou, Organization Science (2024) · Senate "Deadly Denial" report (Nov 2024) · Independent dual-method analysis (Feb 2026) · WEF Future of Jobs Report (2025) · Gartner workforce forecasts (2025) · Sector exposure estimates per Synthetic Outlaw methodology, J. Gropper (forthcoming); deployment-level SO risk scores are published in the Index · Political power index is a qualitative editorial assessment, not a measured variable. All displacement figures are estimates.