Cohort forming

Expungement Eligibility Screening

Help people with criminal records determine whether they qualify for expungement, sealing, or other record-clearing remedies under their state's laws.

← All Workflows

The Package

This workflow is earlier in development than others. A working prototype exists for one state, and the cohort is extracting the reusable patterns. Most assets are still being shaped by the working group.

Working draft

Functional Agenda

What an expungement screening system needs to do: accept a criminal history, apply statutory eligibility rules to each charge, account for how charges interact with each other, explain the reasoning, and generate next steps for eligible and ineligible charges.

Join the cohort to contribute →
Working draft

Conformance Standards

How to test whether a screening system gets the right answer. Because this is a rules-based workflow, evaluation is primarily deterministic: run test cases through the system and check whether the per-charge eligibility determinations are correct. Key dimensions include charge interaction handling and explanation quality.

Join the cohort to contribute →
Working draft

Reference Architecture

The four-component pattern emerging from initial implementation work: a records pull (connecting to court systems), a rules engine (deterministic statutory logic), a charge classifier (the one place AI makes a judgment call), and a document assembler (generating petitions from results). Which parts are reusable across states and which must be rebuilt per jurisdiction.

Join the cohort to contribute →
Planned

Implementations

Case studies from teams building expungement screening systems. What they chose, how they modeled the statutory logic, how they handle multi-charge interactions, and what they learned about maintaining a rules engine as statutes change.

Coming as implementations mature
Planned

Evaluation and Test Suite

A bank of test cases with known-correct eligibility determinations, covering straightforward single-charge cases, multi-charge interactions, edge cases around waiting periods and outstanding fines, and ambiguous situations that should be flagged for attorney review.

Coming as standards mature
Planned

Masterclass

A course for teams building or adopting expungement screening tools. Covers the rules-first design principle, how to model statutory logic as code, how to handle jurisdiction-specific requirements, and how to evaluate a system before trusting it with real clients.

Coming as assets mature

What This Workflow Is

Across the country, millions of people carry criminal records that affect their ability to find housing, employment, and long-term stability. Many are eligible for some form of expungement, sealing, or record-clearing remedy. The vast majority never receive it.

The problem is not a lack of legal authority. Most states have expungement statutes. The problem is complexity. To determine whether a single charge is eligible, an attorney must gather the client's complete criminal history across multiple courts, counties, and agencies. Then they apply a tangle of statutory rules to each charge, accounting for how case sequencing, charge types, fines, fees, and waiting periods interact. For a client with multiple charges across multiple counties, the analysis can take hours.

An expungement eligibility screener takes in a person's criminal history and returns a per-charge determination: expungeable, not expungeable, or flagged for attorney review. Each determination comes with an explanation of the reasoning, the relevant statutory provisions, and concrete next steps.

Who this serves: Legal aid reentry attorneys, public defender offices, Clean Slate implementation teams, community reentry organizations, and any program that screens people for record-clearing eligibility.

Core design principle: if it can be avoided, don't use AI. Expungement eligibility is a rules-based problem. The statutes define clear conditions. The screening system should be built as deterministic code that models the statutes as a decision flowchart, not as LLM prompts. AI is useful for extracting data from messy documents and classifying free-text charge descriptions. It should not be making eligibility determinations.

Join This Cohort

The Expungement Eligibility working group is developing the shared assets on this page. The group is open to reentry attorneys, public defenders, court IT teams, Clean Slate implementation teams, researchers, and technology builders. Join to help shape the functional agenda and conformance standards from the ground up.

Join this Working Group See All Working Groups