Robot Litigation Lawyers — Part 1: A Justice API
Future or fiction?
Despite three months of automated chasing, Amy’s customer had still not paid her invoice. She got out her phone and started a conversation with a robot lawyer. It asked her questions about the debt and she explained what had happened. She uploaded documents including the original contract. Finally, the robot concluded that her case had a high chance of success and asked, “Would you like me to issue the claim for you?”
“Yes,” she replied. The robot sent the claim details and supporting evidence to the relevant court. She received an email and mobile notification confirming receipt. A week later the robot notified her that a defence had been filed and started a conversation to help her prepare for the hearing.
This scenario isn’t as remote as it might seem. In fact, most of the elements are beginning to fall into place:
- Accounting software such as Chaser can automatically chase invoices
- AI-powered platforms (“robot lawyers”) such as DoNotPay and LISA are already providing legal advice through chat
- Tools like those from LegalSifter and Seal Software can analyse contracts and extract key data
- Lex Machina and other legal analytics products can provide the data needed to predict case success
- Solutions like Avvoka can automatically generate documents like court filings
Accessing justice by API
The one missing piece of the puzzle is a way for the robot to interact directly with the court. A human lawyer does not just advise a client, but also acts on their behalf in court. If robot lawyers are to evolve to take the place of human litigation lawyers, even in simpler matters like debt-recovery, they must be able to take action for their clients in the same way.
Enter APIs. APIs (‘Application Programming Interfaces’) allow separate computer systems to communicate with each other in a pre-defined format. In my example, the robot lawyer might send the details of the parties, the cause of action, and value of the claim to the court electronically. It could expect to receive back a claim number, proof of issue, and timetable for the case. All of this would take place behind the scenes, with the robot lawyer keeping its client informed and taking instructions.
Is a justice API feasible?
Technically, this is well-trodden ground. Elsewhere in government, HMRC already has dozens of APIs online which are accessible to third party developers. As a result, taxpayers can file tax returns directly from their accounting product with a single click. The software automatically gathers the relevant information and sends it electronically to HMRC. These APIs also allow applications to access information on users’ tax status and entitlements. This provides, in principle, some of the information a robot lawyer might need to give tax advice.
HM Courts & Tribunals Service (‘HMCTS’) is currently engaged in a £1bn modernization programme. The programme includes new systems for interacting with litigants, including the Online Divorce Service and revamped Money Claim Online service for small civil claims. Both of these court services rely on APIs to connect the online forms users can see with the back-end systems which support them.
Unlike HMRC, however, HMCTS has not made its APIs available to third party developers. Even though the APIs exist—you can even see the code they run on!—there is no way to engage with the online court services except through the online forms provided by HMCTS. To allow robot lawyers to act directly for their clients in litigation, HMCTS will need to extend its APIs and make them accessible to third parties.
Where to start?
Developing a comprehensive justice API won’t be easy. Many court processes still run on pen and paper, let alone on the internet. I’ve written previously about where to start when automating tasks through chatbots. The trick is to target the tasks which arise most frequently and will be easiest to automate. For justice, this might be issuing small civil claims. Every claim is issued at some point, and the API for small claims exists already. More complicated and less common tasks like filing counterclaims or making interim applications can come later.
Done right, a justice API would allow developers to build an ecosystem of applications which interact with and add value to the underlying court service. In Part 2, I will explore how these robot lawyers could enable access to justice and how they might be regulated.
William T. White