Regulatory Consulting at Vera Institute of Justice

Overview

The Vera Institute of Justice is a national research and advocacy organization working to transform the criminal legal and immigration systems, and it has become one of the sharpest institutional voices on AI in those systems. Founded in 1961 to advocate for alternatives to money bail in New York City and now headquartered in Brooklyn with offices in Washington, DC, New Orleans, and Los Angeles, Vera partners with impacted communities and government leaders to develop just, antiracist solutions. Its AI-policy contribution is concrete rather than abstract: Vera has developed five AI accountability principles setting out how AI should be deployed safely, transparently, and equitably in criminal justice and immigration, and it argues that agencies should avoid using AI in ways that expand the reach of those systems, heighten public surveillance, or increase police deployment. That work sits alongside Vera’s landmark empirical research on electronic monitoring — including a national census drawing data from all 50 states, more than 800 counties, and the federal court and immigration systems — which documented a nearly fivefold increase in people under electronic surveillance between 2005 and 2021. For a directory focused on AI policy, Vera covers the domain where algorithmic error carries the highest human cost.

AI-Focused Regulatory Work

Vera’s case against unaccountable criminal-justice AI is built on documented harm rather than hypothetical risk. Its researchers point to racial bias in predictive risk-assessment and policing tools trained on crime data, which reinforce historic overpolicing of marginalized communities and magnify racial inequities at scale; license-plate reader errors that falsely flag cars as stolen, leading to innocent people being stopped and arrested at gunpoint; and the opacity of ‘black box’ algorithms making bail, sentencing, and parole recommendations that defense attorneys cannot meaningfully scrutinize or challenge. The institute’s touchstone case is Robert Williams, arrested and jailed after facial-recognition technology wrongly matched grainy surveillance footage to his old driver’s-license photo — handcuffed in front of his young daughters and held overnight for a crime he did not commit. As Vera’s researchers put it, unchecked AI entrenches existing biases and makes the justice system more unjust at scale. The organization’s response is not blanket opposition but structured accountability: five principles for transparent, equitable deployment, plus a clear line that AI should not be used to widen surveillance or system reach. That position lands in a live policy moment, as federal AI policy shifted sharply in 2025 and the Department of Justice issued its own guidance on AI in criminal justice covering pretrial risk assessment and sentencing tools. Paired with Vera’s electronic-monitoring data — including ICE more than tripling adults on electronic monitoring to roughly 360,000 between 2021 and 2022 — the institute offers something rare in AI policy: rigorous numbers about who is actually being surveilled, and what happens when the algorithm is wrong.

Notable AI Policy Work

• Developed five AI accountability principles for deploying AI safely, transparently, and equitably in the criminal justice and immigration systems — including the position that agencies should avoid AI that expands system reach, heightens surveillance, or increases police deployment.
• Documented concrete algorithmic harms: racial bias in predictive policing and risk-assessment tools, license-plate-reader misidentification leading to wrongful arrests, and opaque ‘black box’ bail, sentencing, and parole recommendations that attorneys cannot challenge.
• Published ‘People on Electronic Monitoring’ and ‘Electronic Monitoring Trends in the United States, 2005–2022’ — a national census gathering data from all 50 states, 800+ counties, and the federal court and immigration systems.
• Found electronic monitoring grew nearly fivefold from 2005 to 2021, with ICE more than tripling adults under EM to roughly 360,000 between 2021 and 2022 — evidence that EM often expands rather than replaces incarceration.
• Advances jail-decarceration strategies with government and civic leaders, and highlights jurisdictions that have implemented them successfully.

Thought Leadership

1. AI Use in the Criminal Justice and Immigration Systems Requires Strong Safeguards (2026)
   Sets out Vera’s five AI accountability principles and documents the harms — biased risk tools, license-plate misreads, and black-box bail and sentencing recommendations.
2. Report on the Growth of Electronic Monitoring in the United States
   Documents a nearly fivefold rise in electronic monitoring from 2005 to 2021 and ICE’s tripling of EM placements, challenging the claim that EM is an alternative to incarceration.
3. Vera Institute — Solutions & Research
   Vera’s research hub across policing, pretrial and sentencing, immigration, and the data and technology that increasingly drive those systems.

Clients

Nonprofit; partners with government and civic leaders rather than serving commercial clients. Works with state and local criminal-legal agencies, jurisdictions pursuing jail decarceration, and impacted communities. Headquartered in Brooklyn with offices in Washington, DC, New Orleans, and Los Angeles.