Artificial intelligence and the legal sector – what next?

8 November, 2017

Artificial intelligence continues to make headlines, this time for its application to the legal sector. With wider integration of artificial intelligence and law envisaged, lawyers and their regulatory bodies must grapple with some difficult questions.

Man vs machine

Recently, legal technology firm CaseCrunch invited lawyers from leading London firms to pit their wits against its CaseCruncher Alpha software. The software is a ‘legal decision predictor’ designed to resolve specific problems. The challenge involved each side predicting the outcomes of real, previously-decided payment protection insurance (PPI) mis-selling claims. The result was a clear victory for artificial intelligence: 86.2% of CaseCruncher Alpha’s predictions were accurate, compared to only 62.3% of the lawyers.

These figures suggest that an increasing role for artificial intelligence within the legal services sector would be beneficial. After all, accurate decision-making saves time and money.

A closer look

Despite technological advances, the advent of artificial intelligence in law should not cause too much alarm for lawyers at present.

Firstly, by CaseCrunch’s own admission its software requires precisely defined questions with binary responses. Its victory in this prediction challenge also came against lawyers with little experience of dealing with PPI claims in practice. The software was provided with a database of information to use. In contrast, its human adversaries had to research their materials from scratch in a limited timeframe.

The issues highlighted above suggest that the way forward is for lawyers to work alongside legal artificial intelligence resources. Humans can still conduct the fact finding exercise and fulfil a supervisory role whilst artificial intelligence carries out intensive data-heavy processes. This way, clients retain an empathetic point of contact who continues to provide a personal service.

The regulatory conundrum

Having established that ‘the machines’ are not imminently taking over, we have to consider how the artificial intelligence-lawyer relationship will work. The legal sector is a restricted one, with only regulated individuals permitted to provide certain reserved activities.

The Solicitors Regulation Authority (SRA) maintains a Handbook on code of conduct, but this does not currently reference artificial intelligence. This leaves several important questions unanswered. For example, should clients have a right to know of the role artificial intelligence played in the advice given to them? In litigation, should the use of artificial intelligence be revealed to opponents? Artificial intelligence predictions of case outcomes could have tactical implications in deciding whether to settle or proceed to trial.

Additionally, the SRA’s handbook uses non-binding ‘indicative behaviours’ to guide solicitors. This introduces a level of subjectivity which goes beyond the binary capabilities of software such as CaseCrunch Alpha.

What next?

The forward march of artificial intelligence is inevitable but as lawyers and clients, we have it in our grasp to determine the direction of progress in the legal sector. Until the relationship between law and artificial intelligence is more clearly defined, we must keep regulatory concerns at the forefront of our minds.


For more information on artificial intelligence, please contact Phil Bilney at or on +44 (0)1732 224 046.

For updates on the latest Tech news, follow us on Twitter @CrippsTechLaw