Pro Bono by Algorithms: Rethinking Accessibility of Law in the Chatbot Era
By Irene Ng (Huang Ying)
At the recent ChatbotConf 2017 hosted in Vienna, Austria on October 2-3, distinguished speakers from leading technology companies convened to discuss an up-and-coming tech – none other than the chatbot. The speakers discussed a range of topics, such as “Competing with character”, “turn(ing) conversations into relationships”, and “building conversational experiences”, and other topics, which is viewable at the ChatbotConf website.
If you thought that the above topics were describing human relations, the fact is, you were not exactly wrong – the focus is actually about developing a human character for chatbots. For some of us, the chatbot might be a piece of tech that we are acquainted with. We may have interacted with these bots on social media platforms such as Facebook Messenger, or used bots on Twitter to track down Pokémon to catch on the famous assisted virtual reality game, Pokémon Go. In some cases, these chatbots are designed to provide customer support or service to the target audience. In other cases, such chatbots are built to provide simple, updated information to users, such as the TyranitarBot on Twitter, or Poncho, a bot that is designed to send “fun, personalized weather forecasts every morning”.
This growing prevalence and use of chatbots by businesses or organizations on various platforms is not something to be ignored. Within the legal industry, several companies have created “legal bots” that are designed to either direct users to the right place (e.g. what kind of lawyer they should be seeking), or perform an easy, repetitive service that can be easily automated and resolved. A famous case displaying the potential of chatbots in the legal industry is that of DoNotPay, a chatbot that has reportedly helped “overturn 120,000 parking tickets in New York and London” by challenging parking tickets. Besides DoNotPay, there are other bots in the legal industry such as LegalPundits that helps to determine what kind of legal advice the potential client needs, to “match [the client] with the resources that [the client] needs”.
As users become more comfortable with interacting with chatbots and using chatbots to help them solve their customer queries, an interesting avenue to explore is the use of chatbots for institutions providing pro bono services. Institutions that provide pro bono services, in particular those that run free legal clinics, can benefit from the use of chatbots in various ways. Firstly, these institutions can use chatbots as a screening tool to filter out whether the said applicant has met the means test to qualify for the free pro bono services. Means tests usually require applicants to fulfill a fixed set of criteria, and if such criteria are generally inflexible (e.g. applicant’s income must be less than USD$1,000.00, anything above this amount will be rejected), then the chatbot can be deployed to interact with these applicants to determine whether the applicant has, at the first screening, met the basic criteria for free pro bono services.
Similarly, institutions can use these chatbots to direct applicants or callers to the right ministry or non-profit organization that may be able to assist them further in the specific legal query that they have. For example, an institution providing pro bono services may often get inquirers making simples requests, such as “where can I repeal my parking ticket”, or “how do I get a divorce”. For the latter scenario, the chatbot can be trained to provide a response, indicating that the inquirer ought to seek a divorce lawyer, point the inquirer to a set of easily digestible information on divorces, followed by a list of divorce lawyers that the inquirer may contact.
Granted, there may – or will – be pitfalls in using chatbots to deal with legal pro bono queries. Applicants or inquirers that approach institutions providing pro bono services may become emotional when discussing their legal problems, and having a human touch attending to such a person’s legal needs may seem to be preferable than a machine. Furthermore, while chatbots can be trained to fulfill certain functions such as determining whether an applicant meets the means test, borderline cases may not be adequately attended to. Using the means test example provided earlier, where applicants must have an income of less than USD $1,000.00, an applicant who declares that she earns USD $1,001.00 may be rejected by the chatbot automatically if the developer did not train the chatbot to consider such borderline cases.
However, despite these concerns, there is still much room for chatbots to grow and help serve a public service function by providing greater accessibility to law. A good chatbot can help pro bono institutions make better use of their resources. By implementing a chatbot to help with simple tasks such as diverting inquirers to the right pages, or assisting volunteers to sift out genuine applicants that fulfill the means test, these pro bono institutions can divert resources or manpower, which would otherwise be used to tackle these relatively simple and repetitive tasks, to other areas, thereby increasing efficiency with the same limited budget that such institutions providing pro bono services have.
While there has been much chatter in the chatbot scene to develop an emotional intelligence for chatbots, ultimately, providing legal aid is a form of public service – and as with all types of service, it is unavoidable that humans may still want to converse with a real human being. As we move forward to explore new avenues of providing legal aid through different platforms in a more efficient and cost-effective manner, we should never forget nor neglect to still provide a physical helping hand to those who need legal aid – and not assume that a chatbot can take our place and release us from our social duty as lawyers to help the needy.