Skip to content
/ expunge Public

chatbot to help understand eligibility for criminal-record expungement in Delaware

License

Notifications You must be signed in to change notification settings

dawn-i/expunge

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 

Repository files navigation

About

This is a simple chatbot created for the Delaware Office of Defense Services, a statewide public defender's office with a holistic defense model. It walks users through the complex factors that will determine whether their criminal record is eligible for expungement and leads to, essentially, one of three results: likely eligible now, likely eligible after specific additional conditions (such as after more time has passed or fines have been paid), or likely ineligible.

Interact with it at Can I expunge my criminal record?

See it in the wild at Expungement in Delaware.

Contact

If you are a public defender’s office or justice advocacy group interested in making a similar chatbot for your state, please do not hesitate to contact me. Whether you're merely exploring the idea, are interested in the concerns we addressed with our implementation, have technical questions, need to contract a developer, or just want casual tips to get started on your own, ... I'm happy to chat with you.

Wishing you all the best with your legal justice efforts.

Connect with Dawn I.

Make your own

Here are the general steps to making your own chatbot.

  1. Get a solid grasp on your jurisdiction’s law, policies, and procedures related to the chatbot's topic.
  2. Map the decision information.
    • Simple option: Map the info as linearly as possible. For example, if you put it into a flow chart, you would want most of the operations to be in a straight line, without a lot of branching or loops, from start to finish.
    • Advanced option: Map the info in terms of (a) all the variables you might need to know, (b) all the relevant values, (c) all the calculations you might need to make, and (d) all the logical structural rules. For example, for the variable "offense level", the values might be "felony type C", "felony type B", and "misdemeanor type A" or whatever relevant labels your jurisdiction uses; a calculation for the variable "time passed" might be "current date minus date of conviction, rounded down to nearest year"; and a rule might be "if offense level is felony type C, and if time passed is less than ten years, then current eligibility is false".
  3. Use the map to write interactive questions, responses, explanations, and next steps. Write in plain language, not legalese. Also write any introductory language that your office deems necessary (such as, you know, “This chatbot is not your lawyer.”).
  4. Code your chatbot.
    • Simple option: Use QnA Markup, an amazing tool from Suffolk University Law School's LIT Lab "for people with little or no programming experience. It was designed with attorneys in mind and transforms blocks of text into interactive question and answer sessions".
    • Advanced option: Use docassemble. You've already done the bulk of the work by mapping the logic structure.
  5. Style and design.
  6. Test on your website and debug.
  7. Go live.

Releases

No releases published

Packages

No packages published

Languages