Introduction to algorithms

Introduction to algorithms

2020 has introduced formerly technical jargon into common parlance, largely due to COVID-19, from ‘R rate’ to ‘PPE’. Perhaps one of the most memorable words to be plucked from relative obscurity has been ‘algorithm’ -  when the Prime Minister blamed the exam fiasco on a ‘mutant algorithm’. But exam boards are not unique in their (mis)use of algorithms. Legal technology products routinely use various algorithms to help lawyers with their work. So what exactly are algorithms and how are they used in the legal sector?


What is an algorithm?


According to the Oxford English Dictionary, an algorithm is: “A process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer.” These days, most algorithms are coded into computer software which is then often applied to large sets of data and used to perform a certain function. An example of a fairly simple algorithm is the software used to prompt GPs to send out specific alerts for various health screenings based on age and gender of their patients.

In the case of the exam debacle, the algorithmic software was designed to calculate grades for A-Level students based on a combination of teachers’ predicted grades, a pupil’s ranking compared to their peers, and school performance over three years. The fact that the algorithm was considered a failure had nothing to do with any mutation, but was instead simply down to poor design (in this case more likely as a result of the decisions of policy makers rather than bad programming).


How are algorithms used in legal technology?


Most computer software used in law firms contains algorithms. At a basic level, when a lawyer composes an email message and selects recipients, the email software will often offer suggestions for auto-completion based on a combination of their most regularly or most recently contacted recipients. In this case, an algorithm may decide that when they type “John” in the “To:” field, they are attempting to contact John Smith who is their senior partner whom they email every day, as opposed to John Jones who was a client they dealt with once three years ago.

Far more advanced algorithms are used in conjunction with AI to create predictive coding tools. Predictive coding is a method of technology assisted review used to assess the relevance of high volumes of documents for purposes of electronic disclosure. Instead of a paralegal assessing the relevance of each document, the software uses a combination of keyword search and iterative computer learning to do the same job in a fraction of the time. Complex algorithms are the foundation of predictive coding, using a set of coded instructions to calculate the relevance of each document. The algorithms adapt according to feedback provided by a lawyer managing the software. But although the algorithms can change according to human input, they cannot ‘mutate’ of their own accord.

Another type of complicated algorithmic software designed for the legal sector attempts to predict the likely outcome of litigation based on analysis of case law (eg Lex Machina). In this case the algorithm will calculate the odds of a particular litigant ‘winning’ their case in court by comparing the merits of their claim with the outcomes of similar historical cases.

As lawyers come to rely on various types of software to do their job, they will increasingly be making use of algorithms, even though these are often hidden below the surface. A basic understanding of how these algorithms work can  help to avoid the case of the tail wagging the dog.



Related Articles:
Latest Articles:
About the author:
Alex Heshmaty is a legal copywriter and journalist with a particular interest in legal technology. He runs Legal Words, a legal copywriting and marketing agency.