A good name is hard to clear: a national report of digital expungement applications

39 minute read

Introduction

It is estimated that 70 million Americans have a criminal record.¹ Having a record creates harms, called collateral consequences, after a person completes their contact with the criminal justice system. These consequences hobble an individual’s ability to find work, housing, or government services. For many, expungement is a possible legal remedy to erase a criminal record. However, the expungement process is hard to understand and not accessible to average citizens attempting to handle their own filing. Further, there are not enough legal service providers to meet demand to help everyone who qualifies. Attempting to bridge these gaps, numerous groups and individuals are leveraging technology. This report assesses the experience organizations had developing expungement technology and provides recommendations to improve the proliferation and impact of these tools.

Having an adult criminal record does not necessarily mean that one was found guilty. For example, individuals may have a criminal record if they were arrested, but not charged; charged, but not prosecuted; and prosecuted, but never found guilty.

Even if someone was never found guilty, a criminal record in a federal, state, or private database can prevent a person from obtaining employment, government services, and education. For example, 87 percent of employers conduct background checks on potential employees, and over 60 percent of individuals are left unemployed within a year of release from prison.² For those lucky enough to find work, their earning potential is significantly reduced. The Pew Trusts found that a criminal record reduces an individual’s annual earnings by 40 percent.³ Nationally, criminal records have been estimated to cost the U.S. economy up to $65 billion each year, due to increased unemployment and decreased wages.⁴

Regarding higher education access, a component to breaking cycles of poverty and criminality, one report found that 66 percent of universities collect criminal record information in the admissions process.⁵ This occurs around the country while there is no proven link between a student having a criminal record and an increased risk to campus safety.⁶ If one gets accepted to university with a criminal record, an individual may also be denied federal student aid, including Pell Grants.⁷

These consequences disproportionately plague minority communities. One recent study found that 49 percent of African-American men are arrested by age 23.⁸ This is compared to the population at large, of whom 33 percent are arrested at least once by 23.⁹ Further, having a criminal record is twice as likely to punish black job-seekers as it is white ones.¹⁰

For many, expungement and sealing are potential remedies to the consequences caused by having a criminal record. Depending on the state, these terms mean different things. Broadly speaking, expungement obliterates the record and sealing hides the record from public view (though it may remain accessible by law enforcement or by court order). This process, however, is not do-it-yourself friendly. Obtaining one’s record, filling out the right documents, filing, and presenting in court all add a layer of complexity and difficulty for pro se petitioners.

For many, legal-aid providers help navigate the expungement process, but indigent legal services are chronically unable to meet demand. There is also private legal help, but it is often cost prohibitive and excludes those that are most disadvantaged.

To overcome gaps in access to expungement services, lawyers, advocates, and technologists are creating online tools, or “expungement apps,” to make the expungement process easier to understand and to increase access to legal assistance. Looking at six projects from five jurisdictions,¹¹ this report explores the process of designing and building these tools, creating a real world community around the tool, and evaluates the effectiveness of these tools. The report concludes with seven recommendations to improve this work and its impact.

The types of expungement tools

To frame the discussion, this section introduces the types of expungement tools surveyed in this report and the purpose behind them.

Expungement apps are typically client- or attorney-facing, with some additional specifications beyond that. Broadly speaking, there are two styles of expungement app, those that determine eligibility through a series of questions, and those that automate eligibility determination and populating forms. The first style asks a series of binary questions to help a user determine if a record is eligible for expungement. Once determined, some projects then create a referral for the user to a local legal aid provider, others just provide the determination and nothing more. The second type takes a case number and extracts the relevant information from a public records database to determine if a person’s record is expungeable. In the case of the projects reviewed for this report, they also auto-populate court documents to be filed in court. Below are the six unique projects addressed in this report; the first four represent the apps that ask questions, the last two are the automated variety:

• Expunge.io¹² was launched in January 2014 by Smart Chicago, a non-profit civic technology organization.¹³ This public-facing responsive web-app was conceived at a summer program, co-run with Mikva Challenge, for youth in the Chicago area. While learning about technology, the youth participants broke into subject matter groups that included a juvenile justice section. Introduced to the idea of juvenile expungement, the youth felt that an app was the best way to explain this system to their peers. In partnership with the juvenile expungement desk at Cook County Legal Aid, Expunge.io provides referral assistance for individuals with a juvenile record in Cook County. The tool only addresses juvenile expungement.

• ExpungeMaryland¹⁴ is a web-based tool created in 2014 by Advocates for Children and Youth (ACY) in Baltimore, Maryland, a child policy advocacy organization.¹⁵ In addition to an assessment of an individual’s eligibility for expungement, the app also provides referrals to pro bono legal groups and non-legal service providers helpful to those with criminal justice system contact. The app makes expungement determinations for youth seeking sealing, youth that were charged as adults, adults seeking expungement, and individuals that could receive vacatur¹⁶ for certain types of solicitation and prostitution charges under state and federal law. At the user’s request, those determinations are sent to local legal aid partners.

• My Clean Jacket¹⁷ is an Android, iOS, and web-app created by the Justice and Accountability Center of Louisiana (JAC), an organization that helps individuals in the post-conviction process. An organization of two with a cohort of volunteers, JAC started to build the app in 2013 as a way to improve their reach across the state. It also became an educational tool for attorneys less familiar with the expungement statute. The app creates referrals to JAC itself or to local attorneys that have paid a donation to be named as a referral organization.

• Clean Slate D.C.¹⁸ was born out of a hackathon held by Rebuilding Reentry in late 2014. Rebuilding Reentry is a component of Mission Launch, an organization that helps those reentering society learn entrepreneurial skills. Built with volunteers from Code for D.C., the tool handles adult records and is currently attorney-facing. At the time of writing, the tool was in beta.

• The Expungement Generator is a tool built by Community Legal Services in Philadelphia, Pennsylvania. Launched in 2011, the tool searches the court’s public records database for a person’s file(s), then processes the files’ information to generate a near-ready, editable filing document. According to the app’s creator, the tool cuts document drafting time down from one hour to ten minutes. The tool is used by attorneys and law students in a clinic format and is not public-facing.

• MDExpungement¹⁹ is a public-facing, web-tool built by Matthew Stubenberg, an attorney in Baltimore, Maryland. The beta launched in January 2015. Similar to the Expungement Generator, the project incorporated Maryland’s state criminal record database so that a user inputs a case number and the app makes a determination of whether or not the case is expungeable. If the case is expungeable, then the app auto-populates a document that can be filed with a court.

The purpose behind an expungement app

While all of these projects are similar in that they improve expungement understanding and access, each organization had a different purpose behind their work.

Sonia Marizano, a project coordinator at Smart Chicago, explains that Expunge.io developed as an intermediary between those that needed a juvenile record expunged and the attorneys that could help. As an intermediary, they wanted to get as many people as possible to existing legal services through their portal.²⁰ Since Expunge.io does not make a legal determination and, instead, refers juveniles to legal help, the tool evolved to have fewer questions to shorten the path to connect with an attorney.

ExpungeMaryland was founded with a similar ethos and intent as Expunge.io. However, instead of being built for the individual with a record, it was built for professionals (e.g., social workers, librarians, and parole officers) who regularly came into contact with people with records, but lacked the legal background necessary to help with the expungement process. The tool now, which is public-facing, also has traffic from individual’s seeking to expunge their own records.

Clean Slate D.C., MDExpungement, and the Expungement Generator were attempts to improve attorney workflow. Clean Slate D.C. accomplishes this by making the expungement statute accessible for attorneys that otherwise might not be familiar with the law. MDExpungement and the Expungement Generator diminished the redundancy that it takes to research and fill out expungements, thus saving attorneys’ time.

Designing & building expungement tools

The design of any technology tool involves a mix of hypothesis, validation, and iteration. This section explores the preexisting requirements needed for these tools, the design steps organizations used, and the cost associated with developing these tools.

Requirements of an expungement tool

Based on the models discussed in this report, there are a couple of necessary factors that need to exist before deciding to develop a tool. These requirements include that the target population has access to the Internet and that there is a controlling statute. For those looking to replicate the Expungement Generator model, an accessible and reliable criminal records database is also required.

All six of the projects reviewed for this report built web-based expungement tools, meaning that the Internet is required to access and use the tool. In building their web-based tools for expungement, all projects also assumed that their targeted end-user has the ability to get online. However, the age, income, and digital nativeness of a desired end-user will potentially affect if an online tool is the right choice. Surveying the app’s intended audience from the beginning will help determine this issue.

For organizations looking to replicate any of these projects, it is important that they consider – not only whether a web-based tool is the right path – but also the local statutes and hearing processes.

Some expungement laws are easier to build a tool for than others. Specifically, jurisdictions that apply an objective factor test regarding what type of record is expungeable are ideal. MDExpungement’s Stubenberg points out that statutes allowing for significant judicial discretion, “good person” factor tests, or a requirement to call character witnesses or victims to testify create a subjective decision-making process that is not suited for the structured decision-making that these tools use. An assessment of the statute by an attorney should be sufficient to make a determination.

Projects like the Expungement Generator or MDExpungement require access to a criminal record database. Although Pennsylvania and Maryland both have reliable, public databases, the majority of states do not. In jurisdictions without a database, this model will be difficult to implement. While it is likely possible to make a public records request for adult criminal record information to create a database independently of the courts, this time consuming approach may create more work than it offsets.

Designing for the potential user

When developing a new tool, getting input from potential clients and partners is important. MDExpungement’s Stubenberg did not survey users about their needs or habits in the early stages of his project. He regretted the choice, saying that taking time at the beginning to understand user needs and design “would have saved a lot of time down the road.”

Two approaches at gathering feedback are Smart Chicago’s CUT Group and ExpungeMaryland’s efforts to make legal jargon accessible.

Smart Chicago’s Civic User Testing

While most organizations tested their ideas with community members and lawyers, many were limited by capacity, experience, and funding to thoroughly test their projects. However, Smart Chicago was uniquely positioned to undertake user testing because of their Civic User Testing (CUT) Group, which pays a community of Cook County residents to test civic websites and apps in a structured way.²¹

Smart Chicago wrote about two focus groups they ran and how it informed an update for Expunge.io. From the outset, Smart Chicago wanted answers to the following questions:

• Is Expunge.io easy to use? Do users like the website? Are there any problems/issues using the website?

• What stands out on Expunge.io? What do users remember? Do users have a good understanding of the expungement process after visiting the website?

• Do they want to share information with others?

• How can Expunge.io be improved?

Through their test group, they found that users were hung up on the fact that the expungement filing process cost money and that one had to be 18 or older to file. The app previously included this information, but users felt it needed to be more prominent. Testers felt that if they were sharing the tool with peers, they would not want this information to be “hidden,” which at the time they felt it was. Another recommendation through this process was to include social media buttons on landing pages to increase sharing.

An Expunge.io landing page with no social media buttons. Courtesy of Expunge.io.

ExpungeMaryland & Legal Jargon

The ExpungeMaryland project came up against a challenge that faces many public-facing legal tech projects: how can legal jargon be made accessible to a non-legal audience? One goal of the ExpungeMaryland project was to have the site text written at a seventh grade reading level. However, words like “expungement” and “nolle prosequi” stood in the way of this goal. Further, if a user read from their RAP sheet to answer questions, then legal terms would be their point of reference. To split the difference, ExpungeMaryland created “mouseover” or hover text for the legalistic and Latin terms that could not be replaced with simple English.²² This approach allowed the project to use the legal terms that people would see on their RAP sheets while increasing understanding of those that did not have their RAP sheet handy. Colleagues and attorneys tested language on the site, but ACY did not have the capacity to undertake a community focus group at the time.

An example of hover text. Courtesy of ExpungeMaryland.org.

The cost of making an expungement tool

By and large, organizations built their tools with the help of volunteers; however, those same organizations found that relying on volunteers only got them so far. Clean Slate D.C. reached out to a local Code for America Brigade for help. A University of Maryland Law School student coded ExpungeMaryland. The legal aid attorneys that wanted to use a tool for their own practice and advocacy created the Expungement Generator and MDExpungement.

My Clean Jacket, which was coded by a professor and students at Loyola University New Orleans, estimated that donated developer and attorney time to build the app totaled $30,000. Michael Hollander of the Expungement Generator estimated that he spent 200 hours over the past four years on the project, which includes training others to use the tool. At a developer rate of $150 an hour, this would also put this project at $30,000 total.

While these organizations got their projects started with the help of volunteers, some articulate a need for funding to improve usability, functionality, and access. My Clean Jacket said that they had reached their limit on what they could accomplish with free labor, and that if they were going to include a form builder or any other feature they would need a grant of $35,000 to cover those development costs. ExpungeMaryland and Clean Slate D.C. also mentioned the lack of funding as one reason that their projects are slow to move forward.

Organizations said that the other costs to setting up and maintaining their projects were low. Excluding maintenance, MDExpungement had the most expensive yearly overhead at $300 a year. Other projects were spending less than $50 a year, primarily on domain names and hosting.

Building community around new technology & improving user adoption

Being an online tool underpins all of these projects; however, real world organization and community outreach are important to project success. As Adrienne Wheeler, the executive director at JAC, put it, “It’s not about just building an app.” Merely building a tool and putting it online does not mean that the intended audience will find it or that people will use it. In this regard, more traditional coalition building and search engine optimization improved people’s access to and use of the tool.

Building a real world community around a new tool

Organizations took different approaches to driving traffic to their platforms, which included direct-to-consumer advertising, reaching out to non-traditional partners, using data to better understand a community’s need, and incorporating the tool into a third-party’s workflow.

Community building is a strong aspect of Expunge.io. Marizano says that the youth driven project took dozens of professionals and a half dozen organizations from different fields to come together and build the project. Through this effort, however, a community cropped up around the tool, which helped create awareness and drive traffic.

On account of this input, Smart Chicago varied their approaches to drive traffic to Expunge.io. First, they received a grant that allowed them to put ads on city buses. Second, they put a kiosk at the Cook County Legal Aid juvenile expungement helpdesk that lets people access the tool. Third, they train youth volunteers to do “lunchroom take overs” in Chicago public schools, which allows for peer-to-peer education about expungement. Fourth, Smart Chicago, with their partners, has educational events where they hand out cards that help draw people to the site. Similarly, JAC sent postcards out to potential beneficiaries of My Clean Jacket.

To build a broad community coalition, ExpungeMaryland and MDExpungement undertook significant community outreach. Shortly after launch, ExpungeMaryland reached out to community organizations, the faith community, librarians, and elected officials to introduce the tool. Elected officials found the tool particularly useful for their constituent services. This increased the size and type of the community supporting the tool, as well as increased the network of non-legal referral organizations the tool could direct users to. Further, it increased traffic to the site. ExpungeMaryland saw traffic grow monthly for the first year after its launch.

To increase knowledge and information about expungement in Baltimore, MDExpungement mapped where all of the individuals with expungeable records are in the city. Beyond visualizing the 937,787 potential expungements in a city of 620,000,²³ this visualization better informs providers where expungement clinics should be held. It also provides an opportunity to reach out to individuals with records that live near an upcoming clinic.

Further, MDExpungement became a part of the intake process at Maryland Volunteer Lawyers Service (MVLS), a non-profit legal service provider. When a new client comes in for legal assistance the client’s name is put in the app to see if he has an expungeable criminal record. Wrapping this platform into MVLS’s intake process is improving the tool’s impact. According to the app’s creator, one in four MVLS clients has an expungeable criminal record.

Helping potential users find tools online

For public-facing tools, being high on a Google result is one why to improve connection to potential users. This is because fewer than five percent of online searches go beyond the first page of results.²⁴ Expunge.io, ExpungeMaryland, and MDExpungement, spent time on making their sites findable through search engines. This process, called search engine optimization (SEO), started out with the selection of the name. All three use “expunge” in some capacity, and the two Maryland projects also included the state’s name. This is useful because it brands the site in a way that reflects what people search for. For example, see the Google Trend analysis of the word “expungement” search from a Maryland IP address shown below.

Courtesy of Google.com/trends.

The column on the right represents the top five permutations of “expungement” used by those searching in Maryland. The list makes clear that when people are searching for expungement help, they are doing it for Maryland specifically, thus why the literal branding of ExpungeMaryland and MDExpungement has utility. If the name of the website reflects what people are searching for, people are more likely to find it. Through their SEO efforts, ExpungeMaryland and MDExpungement are the third and fifth Google result, respectively, when searching for “Expungement Maryland”.

Regarding online access, the CUT Group taught Expunge.io’s team that two URLs they used, http://www.expunge.io/adultrecord and http://www.expunge.io/adultorjuv, were blocked by parental controls because of the word “adult”. By changing the URL, this problem was overcome and parental controls no longer kept users from finding the site. This was a unique problem for their audience because many users were searching from school computers with parental controls.

Bridging the worlds of SEO and real world community organizing, it is also helpful to have popular Google results for “expungement” link to the project’s website. To this end, ExpungeMaryland contacted the top ten Google results for “expungement Maryland” and asked sites’ owners to include ExpungeMaryland.org as a resource. This helped pull the then nascent project off the back pages of Google results.

The last point on SEO is Google AdWords Grants. This is a grant through Google.org for non-profits that provides funding for Google AdWords, which will promote the site at the top of relevant Google searches. It is simple to apply for, and it significantly helped boost ExpungeMaryland’s place in Google results.

Evaluating impact & improving longevity

While the ingenuity going into increasing access to expungement is impressive, the results are mixed. This section looks to quantify the impact these tools are having, understand the hurdles organizations had after launch, and determine if these tools can offset their costs.

The impact these tools are having

It is still hard to quantifiably evaluate these tools across jurisdictions because some projects are using analytics and some are not. Those that are collecting data are not necessarily collecting the same data, making comparisons challenging.

This is the data that was made available for this report:

• Expungement Generator: This project has the most thorough data on impact and outcome of any project in this report. The creator of the platform says that the tool has been used for nearly 10,000 expungement filings as of fall 2015. Ninety-eight percent have been granted statewide. Philadelphia Lawyers for Social Equity also uses the tool to help “tougher clients” with more extensive criminal records. Their success rate of attaining expungements through the tool is 94 percent. In 2010, the year before the app was first used, the First Judicial District in Philadelphia saw 4,430 expungements filed. The year after the app’s release, 2012, the court saw expungement filings jump to 8,125; in 2013, there were 6,003. The First Judicial District in Philadelphia said publicly that they believe that the Expungement Generator made a noticeable impact on the number of expungements being filed.²⁵

• MDExpungement: While this project launched in January of 2015 as a beta test, it remained under the radar until October 2015. That being said, the project saw 1,000 unique petitions printed between January and October. Between October 2015 and February 2016, when the site was fully public, 6,400 unique petitions were printed. The project’s creator says that the tool is primarily being used at expungement clinics in Baltimore City and by attorneys that file numerous petitions.²⁶ Since anyone can print these petitions, it is not currently possible to know how many have been filed or granted.  The project’s creator envisions a feedback loop with the court’s public website that updates as the state removes records from the database.

• Expunge.io: This project made their click through rate available for this report. Of the users that came to the site’s homepage, 40 percent clicked to see if they were eligible for expungement, but only twelve percent got through the disclaimer, which raised concerns that there are too many steps between start and finish. They are unable to track the number of filings that occur on account of the app.

• ExpungeMaryland: While the tool has seen steady traffic, upwards of 1,200 unique visitors a month, this has not translated to significant filings. Statewide the platform has led to 124 filings total, all but two are pending at the time of writing.

• My Clean Jacket: Fully launched in January 2014, the platform has had 750 individual users across the country as of December 2015. Forty of those users have clicked through for a referral. They have not tracked actual filings, because filing has primarily happened through their attorney partners.

• Clean Slate D.C.: No data available at the time of writing.

Hurdles organizations experienced after the initial launch

After launch, projects experienced numerous challenges, including data collection regarding project referrals and outcomes, understanding the number of people who could potentially receive an expungement, and succession planning in the case of statutory and organizational changes.

Expunge.io, which is collecting data through the site, struggles to track who physically comes to the legal aid help desk because of the app. A part of this issue is because two legal organizations staff the help desk and they do not share information between themselves. The project has not figured out how to accurately track case outcomes.

ExpungeMaryland, which includes a list of non-legal referral services for end users, does not track click-through rates of those links. Similarly, the referral organizations do not know if the referral is coming from ExpungeMaryland or somewhere else. This makes it impossible to know if users are using this information.²⁷

All of the projects surveyed want to see more traffic to help more people expunge their criminal records. However, most projects do not know how many potential expungements exist in their jurisdiction, which means there is no baseline to judge if traffic is good or bad. For example, the impact of 100 monthly referrals creates a different impact in a jurisdiction that has 10,000 as compared to one million potential expungeable records. As mentioned above, MDExpungement and the Expungement Generator assessed the public records database to get a sense of how many expungeable records there are in Baltimore and Pennsylvania respectively. However, no other project knows what “good” traffic would look like.

Another post-launch challenge that creators face is keeping the app legally up to date as statutes change. Less than six months after ExpungeMaryland launched, the state created juvenile expungement for the first time. A year after that, the expungement law was reformed significantly, which increased the type of cases that were expungeable. However, at the time of writing this report, the project still reflected the 2014 law. Since the launch of the project, the creator and developer of ExpungeMaryland left the organization, and as current leadership states, the project is not their “highest priority.” Still, the project is live and people will find it high on Google results for expungement in Maryland, which means that individuals may be told that they cannot expunge their record when in fact they can. ACY says they struggle to find a developer to update the project.

Charging a fee to support the tool

Only MDExpungement and My Clean Jacket experimented with a pay structure. MDExpungement started with a $10 user fee. Stubenberg says this was to offset his costs to keep the project alive. After just five expungements, he decided to make the tool free. There is no fee for attorneys that use the tool to help clients.

My Clean Jacket, owned and operated by a non-profit, intends to be financially self-sufficient. To this end, attorneys can donate money to the project and become named referrals through the app. Currently, the fee structure is $400 for a six month listing and $600 for a year-long listing. Ten attorneys statewide have made the donation as of January

1. In the next iteration of the project, Wheeler hopes to include a form builder feature that will have a subscription model for attorneys, while pro se litigants will be able to use it for free.

Recommendations

Understanding the experiences of these six projects provides insight into where this work has yet to go. Based on feedback from these projects and independent analysis by SIMLab, we make seven recommendations to improve the impact of expungement tools.

1. Expand understanding of the greater expungement system.

Either as referral tools or form generators, these tools are dependent on larger legal and bureaucratic systems. To understand this system better, projects should map the broader expungement ecosystem. In doing so, jurisdictions can better understand all of a bureaucratic system’s moving parts as they pertain to expungement, better visualize potential roadblocks, and more effectively propagate solutions.

One potential roadblock affecting the impact and development of expungement tools is access to accurate criminal records. States like Maryland and Pennsylvania have reasonably accurate public record databases for the use in these projects. However, similar databases exist in a minority of states, and so other steps must be thought of to overcome this disconnect. Having limited access to records increases the time expungement takes and the chance that a user disengages with the process.

Expunge.io provides a good example on how to overcome this roadblock. When Expunge.io first launched, a client would be referred to the legal aid help desk via the app, then the client would go to the police department to request a copy of his RAP sheet, which then took two weeks to arrive. Expunge.io and their partners went through great efforts to gain access to the CLEAR police records system to look up RAP sheets at the help desk. While a great step forward in streamlining this process, it does not overcome getting records from outside Chicago.

2. Quantify the need for expungement.

A part of understanding the larger ecosystem is quantifying the number of potentially expungeable records in a jurisdiction. Of the 70 million Americans with a criminal record, only a fraction of them will be eligible for expungement. Only MDExpungement and the Expungement Generator took this step. Where this information is difficult to attain, one potential way to get a ballpark figure is to understand case outcome rates in a particular jurisdiction. For adult records, charges and case outcomes should be public record. The total number of potentially expungeable outcomes less the number of open cases would provide a rough sense of the need. This step will be useful for non-profit organizations to show the size and scope of the problem and convince funders that this is a community-wide issue worth tackling and funding.

3. Improve data collection and case tracking.

Each project wanted to improve its data collection ability. Only one project, Expungement Generator, tracks outcomes. To understand the capacity of these tools, collecting outcomes of filings is critical. Accomplishing this will look different depending on the jurisdiction and the project’s partners. For jurisdictions that cannot automate their data collection through a public records search, it would be valuable to create a feedback loop with either end users or partner attorneys to capture this data. This feedback loop would also create a check for the bureaucracy to make sure that it is thoroughly and quickly executing expungements after they are granted.

Data collection should also be deployed in regards to understanding referral services. While ExpungeMaryland has a resource page for non-legal services, there is no current way to capture the click-through rate or to follow up with those service providers to understand how many referrals come through the site. This coupled with qualitative user feedback would improve the utility of these referrals and inform how to better tailor the page for users.

Better data collection could also improve legislative advocacy. Both ExpungeMaryland and My Clean Jacket have designs to use the site’s data to inform the policy debate around expungement. While neither has actualized this plan, the idea is to quantify which part of the statute is most likely to preclude someone from expungement by tracking click-through rates. By having this information, advocates can be better informed about what parts of the expungement statute, if changed, would provide the greatest impact. Taking this step is one way to increase a tool’s impact in the greater expungement ecosystem as discussed in recommendation one.

4. Explore new ways to improve and increase automation.

Judging by the numbers made available for this report, more workflow automation correlates with more expungements filed. This is likely because it improves the efficiency of creating appropriate documents by cutting down on redundancy. Below are examples of different types of automation or automation workarounds and ideas for where automation can head.

One example of increasing automation is the Expungement Generator’s recent iteration. Previously, when an attorney used the Generator they searched the court records database manually, downloaded the relevant documents, and then uploaded those documents to the Expungement Generator. Now, the app automatically searches the court records database at the attorney’s request, saving time.

Where automation was not possible, the developer of the Expungement Generator created a visualization tool to compliment the attorney’s work. A new law in Pennsylvania allows for sealing of certain second and third degree misdemeanors. On account of this change, the Generator now identifies sealable cases as well as expungeable ones. One challenge in automating the search for sealable records is that the court records database frequently does not mention the degree of the charge. To overcome this, Hollander created the database visualization shown below. The new search engine allows an attorney to enter a criminal statute and see the historic application of the charge over the last five years and whether a particular charge is potentially sealable under the new law. The lines in red are not sealable, the ones in white are.

Beyond this specific work, it is also worth exploring how automated the entirety of the expungement process can become. Specifically, does a person with a criminal record have to be involved at all? It is not out of the realm of possibility to build an expungement bot that would search a criminal record database, create the appropriate documents to file for expungement, print every expungement the bot could find, and then an attorney would file them with the appropriate court. This model raises interesting questions around cost, whether or not this would constitute representation of a client, what happens if there is an error, and how to get an individual’s signature, if needed, for a filing. The answers to these questions would be unique to the jurisdiction and dispositive on whether or not this idea was worth pursuing. In the alternative, an expungement bot could find a defined class of people for a class action suit to expunge certain records in bulk, like what is being done in Tennessee.²⁸

This hypothetical approach begs a larger question: why do statutes not allow for automatic expungement for certain outcomes? There is legislative movement to increase the type of outcomes that receive automatic expungement. The bulk, automated filing approach could embrace Smart Chicago’s “flood the box” philosophy to bring legislative attention to the fact that people with outcomes found in their favor still have to file for expungement.

5. Build tools that compliment, not burden, providers.

Based on the data collected for this report, the most impactful tools, as measured by number of expungements filed, are those that compliment legal providers (the Expungement Generator and MDExpungement). Both projects dovetail into an existing expungement workflow and improve attorney drafting time. In comparison, Clean Slate D.C. said that their initial legal partner was over burdened with work before the creation of the project. Depending on how the tool is used it could bring in more clients, compounding the problem without improvements to attorney workflow. In Louisiana, JAC found that creating the tool demonstrated their expungement expertise and had the unintended consequence of lawyers sending them potential expungement clients, which, at higher rates, could overload the organization.

Another dimension to this issue is project-organization fit. Specifically, while ExpungeMaryland was developed by and for ACY, a staff change made the tool no longer a priority. The current director of the project stated that the project was not his focus. This leaves the tool in purgatory; while it still remains high on search engine results, its inaccuracy and lack of attention diminish its impact. Succession planning and off-boarding should be considered when developing these projects to improve longevity and impact.

6. Tools should provide a path to relevant non-legal services.

While these tools are focused on the issue of expungement, the referral mechanism should include relevant non-expungement services. For those with criminal records, housing, educational, vocational, substance abuse, and family services are all potentially relevant to their needs. ExpungeMaryland acknowledged the need for this approach, but does not help the user find the appropriate service for their specific need. Including triage questions around these related issues would embrace a “wrap around” service approach, and provide users with a more holistic set of services for their specific needs.

7. Explore ways to impact non-government criminal record websites.

The projects looked at in this report only handle expungement of official records held by government agencies. “Mug shot” websites and other non-governmental websites mean that a person’s criminal record can live beyond an expungement. Even though Google changed its algorithm to suppress “mug shot” websites,²⁹ it can require thousands of dollars to have mug shots and criminal records permanently taken off these sites. Because of the use of off-shore servers, it will be hard to regulate these websites. Expungement projects should explore ways to automate takedown notices, which could include working with police departments to use the Digital Millennium Copyright Act to have these photos taken down.³⁰ This would recognize and potentially impact broader issues affecting those with criminal records.

Conclusion

Collectively, the projects discussed in this report have the potential to increase attorney efficiency, improve individual access to justice, and promote greater awareness to this underrepresented issue. There is, however, room for experimentation and growth as these projects become more common. Doing so will hone the scope and improve the potential of this work.

Appendix: Interviews

Clean Slate D.C.

• Briane Knight, attorney and developer, Washington D.C.
• Laurin Hodge, Mission Launch, Washington D.C.

Expunge.io

• Cathy Deng, DataMade, Chicago, Illinois.
• Chris Rudd, Stanford Design School, Palo Alto, California.
• Sonja Marziano, Smart Chicago Collaborative, Chicago, Illinois.

ExpungeMaryland

• Michelle Swift, Maryland Volunteers Lawyer Service, Baltimore, Maryland.
• Rais Akbar, Advocates for Children and Youth, Baltimore, Maryland.
• Sandy Brown, Mid-Shore Pro Bono Inc., Easton, Maryland.

Expungement Generator

• Michael Hollander, Community Legal Services, Philadelphia, Pennsylvania.

MDExpungement

• Matthew Stubenberg, Maryland Legal Apps LLC, Baltimore, Maryland.

My Clean Jacket

• Adrienne Wheeler, Justice and Accountability Center of Louisiana, New Orleans, Louisiana.
• Judson Mitchell, Loyal University New Orleans, New Orleans, Louisiana.

Footnotes