The legal profession is one of the areas of work predicted to feel a significant impact from the use of artificial intelligence or “AI”. The use of intelligent algorithms that, having “learned” on large datasets, can classify, analyse, and make predictions on data has obvious entry level use in the legal sphere. Reviewing large volumes of data as part of disclosure in litigation or due diligence on corporate transactions is work that the legal world is already comfortable with outsourcing.
Machine learning, seeking patterns in datasets and drawing conclusions from them is the next logical step. This functionality is not simply searching and indexing data but also interpreting patterns within that data. This can include, where natural language processing is concerned, identifying language changes. This might for example flag up the fraudulent use of an email account by a third party. The use of AI for these tasks saves time and money and improves accuracy. It is said to also reduce bias. However, the monitoring of bias is an essential part of using AI. If the data fed into the process is biased the AI will naturally adopt that bias. Emphasis on checking for bias in any conclusions drawn by machine learning must be maintained if this technology is to be used for client benefit.
The ability of AI to research and collate relevant caselaw is also a cost-effective use of the technology. Whilst AI can draw conclusions from its research, most clients would rather have the interpretation carried out by a skilled and experienced Solicitor.
26 % of UK law firms currently make use of some aspect of AI in their practise. An AI process is now routinely used for electronic disclosure for example. The use of AI is now spreading to the courts. Judicial Guidance on the use of AI was issued at the end of last year. It emphasises the need for Judges to observe the highest standards of data protection and security when outsourcing research functions or using third party functionality. It also highlights the need for monitoring potential bias in the conclusions or results drawn from the use of AI.
The Law Society’s Rules Committee has recently implemented a decision to use a specific algorithm to calculate staged payments of judgment debts. This has been described as the use of AI by the courts. However, the considerable derision expressed at this description by those in the profession, let alone those using cutting edge AI in industry, is perhaps an indicator that determining formulae is not what the public thinks of as AI. The adoption of AI to speed up court processes would be most welcome but we are not there yet.
There is no sign as yet of AI replacing the strategic and commercial thinking of experienced Solicitors. There is however a streamlining of time and cost available from the use of AI in large data processes and this is to the benefit of clients. In answer to the question at the start of this article, “where are we?”, the answer is, at the beginning of adoption of this technology.
