The most common question asked by clients is, "what are the chances of winning the case?" The lawyer's answer is informed by a range of factors, including the litigator's accumulated knowledge and experience of other cases. As AI becomes more sophisticated, there is a question as to whether AI can replicate a litigator's expertise, by analysing historic litigation and outcomes, and applying it to the new case.

There are already some AI tools that are making advances in this area. An example is Lex Machina which claims to predict the behaviour of courts, judges, lawyers and parties. Their website suggests that the database behind the AI is comprised of publicly available court records which are updated every 24 hours. These kind of AI tools have the potential to reduce risk in litigation strategy by providing data-driven insights into the likelihood of a case succeeding at trial, which in turn helps a client to decide whether to initiate or defend proceedings and whether settlement should be pursued.

There are, however, some limitations with the current AI models. First, generative AI takes a statistical approach to analysing historical court data. Its answers would be informed only by what has happened before and predictions based on available data do not necessarily give the most accurate answers. How then, would AI comment successfully on a wholly novel issue of law or situation? Secondly, the databases behind AI tools seem to be based on narrow data sets: publicly available court data will not contain information on disputes that do not make it to court, nor commentary on court decisions that show a particular decision is controversial and under scrutiny by the legal community or even Parliament. It is also difficult to see how an AI database could capture lawyers' experiences of appearing in front of specific judges – a factor that often feeds into the assessment of litigation risk and the human aspect of presenting the case in a way which will appeal to a particular judge. In this way, AI datasets do not reflect a litigator's breadth of knowledge accumulated from a variety of sources and experience.

These limitations may be overcome in the future. AI development is at an early stage and as it becomes more sophisticated with the potential to access a broader dataset that goes beyond publicly available court records, it will become more useful in predicting case outcomes and informing strategic decisions. It's possible that law firms might look to create their own information pot based on their litigation team's experiences and use it to complement any external information. The judiciary too may at some point in the future use their AI predictive tools to assist in their decision making. It is very much a 'watch this space'.

Should you require any further information on the issues covered in this article, please contact one of our Disputes and Investigations team.