Techy students defend their lawbot
Last week, the Cambridge law students behind lawbot service Elexirr were called “arrogant” by top solicitor David Greene. Speaking to Legal Cheek, Edwin Coe partner Greene stressed that law is “a people to people business”, and that case-predicting computers like Elexirr lack fundamental, human-like qualities. Now, we hear a response to Greene from Ludwig Bull, managing director of Elexirr.
First, an admission: the impact of artificial intelligence (AI) on the legal profession is generally overestimated. This hype is often damaging to valuable innovation.
Second, a confession: English law is one of the most complex and fascinating subjects we have found. Whatever the Elexirr team can do to better understand the law and contribute to its well-being makes us happy. Our core concern is to understand the law as it is, not make it something else.
Third, and more importantly, some specifications.
Being a lawyer is a very complex job composed of many different tasks. We believe that the better way to approach the impact of AI on the legal profession is to take a piecemeal approach. Some highly patterned work may be easily automated (think of software scanning thousands of documents for inconsistencies in minutes at little cost rather than an associate in months for £££.)
Some highly intuitive work may not be automated in the near future, or ever (as much as we love our bot, Elexis, we would prefer a human being to negotiate on our behalf.)
Another category of the interaction between AI and the law (and the one we are most interested in) is where computers help us understand something we thought we already knew. Is it possible to scientifically predict the outcome of a case? Lawyers have different opinions about what the outcome of a particular case should be. Lawyers even have different opinions about whether there is an exact answer to the outcome of a case. We think there is an objective dimension to legal knowledge and that our AI can prove it.
Fourth, and most importantly, a vision: we think that data science helps us show that there is an objective dimension to legal knowledge because it allows us to represent legal information in a non-conceptual way.
Instead of reading a document about the law and using our understanding of language to figure out which concepts the different words represent and how they interact, we feed the document to our algorithm to show us a mathematical representation of the information. We then use that mathematical representation to find patterns and gain insights invisible to a linguistic approach (have you ever tried engineering your own data structure for thousands of Dutch insurance claims to understand whether the emotional state of the claimant affects the outcome? We think not…) Our products are not only a useful tool for consumers and corporations — we think they also shed light on law’s nature.
Fifth, a conclusion: machines will not replace lawyers and Elexirr isn’t trying to change that. Machines can help lawyers understand the law and maybe even make it clearer and more just. Elexirr is a small company trying to take giant leaps. We would be grateful for the support of Mr Greene and other senior members of the legal profession who enjoy our sincere respect. We would also like to extend an invitation to Mr Greene to challenge our system during the competition. Ultimately however, what makes us happiest is to understand and solve law.
Ludwig Bull is a law student at the University of Cambridge and the managing director of Elexirr.
For all the latest commercial awareness info, and advance notification of Legal Cheek’s careers events, sign up to the Legal Cheek Hub.