Skip down to main content

OII Colloquia: What’s wrong with using machine learning to personalise law?

25 Jan 2018

In the legal services industry, automation and the power of machine learning has been transforming legal practice, by automating tasks that had previously been highly labour intensive and typically undertaken by junior lawyers, such as document management, legal discovery and due diligence.  Although there is increasing concern about how these digital transformations will affect  the job market for lawyers, and for legal services more generally, less  attention has been given to the way in which machine learning might be used to ‘personalise’ the law.  In particular, several legal scholars claim that the data-driven techniques used by commercial providers such as Amazon and Google to personalise the user’s experience (whether in the provision of shopping recommendations or responses to search queries) could also be used to ‘personalise’ the law.   In this paper, we consider the claims made by those who believe that such a move would enhance the administration of contemporary legal systems.  In so doing, we offer our preliminary thoughts, suggesting that these proposals are not merely morally problematic, but – if implemented – might fundamentally threaten the rule of law upon which the legitimacy of contemporary legal systems rests.